CN114238608B

CN114238608B - Intelligent interview system and method

Info

Publication number: CN114238608B
Application number: CN202111554807.6A
Authority: CN
Inventors: 杜海涛; 何龙; 杨羽
Original assignee: Beijing Doumi Youpin Technology Development Co ltd
Current assignee: Beijing Doumi Youpin Technology Development Co ltd
Priority date: 2021-12-17
Filing date: 2021-12-17
Publication date: 2023-03-24
Anticipated expiration: 2041-12-17
Also published as: CN114238608A

Abstract

The invention discloses an intelligent interview system and method. Wherein, this system includes: the classification evaluation module is configured to perform classification operation on the expectation of the user through a pre-established expectation analysis model to obtain an expected analysis result, and perform capability level classification evaluation on the capability of the user through a pre-established capability level analysis model to obtain a capability analysis result; a dispatch output module configured to match the user based on the desired analysis results and the capability analysis results; the interaction calculation module is configured to generate a pre-interaction scheme based on the relevant information of the matched users, and analyze interaction data generated by the matched users in the interaction process to generate an interaction response scheme; and the simulation interaction module is used for creating a virtual character based on the pre-interaction scheme and generating the response of the virtual character in the interaction process based on the interaction response scheme so as to carry out intelligent interview.

Description

Intelligent interview system and method

Technical Field

The invention relates to the field of AI intelligence, in particular to an intelligent interview system and an intelligent interview method.

Background

In the recruitment and interview process of the human resource industry, the interview time and the interview place can not be realized. Sometimes, the interview time of the recruiter cannot be matched with the job seeker, or the interview place is far away and the traffic is inconvenient, and the like.

The interviewing is also enormous for the consumption of manpower, for example, a factory has a lot of posts to attract people, but the number of interview officers is insufficient, so that the candidates cannot be interviewed quickly and efficiently and can be effectively fed back. The recruitment period is long, and if only one recruitment is carried out, manpower and material resources are invested in a large scale, and when the large-scale recruitment is not needed any more in the follow-up process, the previous investment is wasted.

Environmental factors may influence the interviewing process, so that the job seeker is nervous or the stress in the mind is too heavy to normally exert. Finally, the work which should be fully qualified is not well performed in the interview, so that the chance is missed.

The interviewer has limited work while interviewing and is prone to fatigue after repeated interviews of multiple candidates. At this time, the judgment and logical thinking ability are greatly reduced. The effectiveness of the interview will be reduced at this time, which results in a failure to evaluate the candidate effectively and correctly at a later time.

In the traditional interview process, a recruiter is required to search and screen resumes and then invite an interview for job seekers. In the searching process, the searching and screening capability of a single person is limited, and because the data amount of personnel in the whole industry is huge, a large amount of resume screening in a short time is difficult. At this time, the intelligent system is required to assist in carrying out corresponding processing.

In the traditional interviewing process, conflicts, contradictions or risks can be generated for various reasons sometimes. And sometimes with unpredictable serious consequences. For example, emotional problems, psychological problems, etc. may cause the job seeker or interviewer to have an overstimulation during the interview process, thereby causing adverse effects.

The traditional interviewing process is not sensitive to the timeliness of the information, and the information is seriously lagged in many times, so that the whole recruitment period is long. For example, a new job seeker has just uploaded resumes, but manual screening and searching takes a long time due to the fact that resumes are many in the system, or the next day is likely to be covered due to the fact that uploading is late at night, and the resumes cannot be viewed at the first time.

The interviewing process of the human resource industry is often influenced by social environment factors, for example, a certain area is blocked due to an emergent social event, so that the interviewing process cannot be carried out if the area is interviewed. At this point, an available interview system is needed to interview the person.

In the traditional interviewing process, the interviewing result analysis often causes the truth of the interviewing result to be deviated due to the subjective impression of people, and further causes job seekers to miss opportunities or enterprises to miss the addition of a good candidate.

In view of the above problems, no effective solution has been proposed.

Disclosure of Invention

The embodiment of the invention provides an intelligent interview system and method, which at least solve the technical problem of low interview efficiency caused by manual interview.

According to an aspect of an embodiment of the present invention, there is provided an intelligent interview system, including: the classification evaluation module is configured to perform classification operation on the expectation of the user through a pre-established expectation analysis model to obtain an expected analysis result, and perform capability level classification evaluation on the capability of the user through a pre-established capability level analysis model to obtain a capability analysis result; a dispatch output module configured to match the user based on the desired analysis results and the capability analysis results; the interaction calculation module is configured to generate a pre-interaction scheme based on the relevant information of the matched users, and analyze interaction data generated by the matched users in the interaction process to generate an interaction response scheme; and the simulation interaction module is used for creating a virtual character based on the pre-interaction scheme and generating the response of the virtual character in the interaction process based on the interaction response scheme so as to carry out intelligent interview.

According to another aspect of the embodiments of the present invention, there is also provided an intelligent interview method, including: carrying out classification operation on expectations of users through a pre-established expectation analysis model to obtain expectation analysis results, and carrying out capacity level classification evaluation on the capacities of the users through a pre-established capacity level analysis model to obtain capacity analysis results; matching the user based on the expected analysis result and the ability analysis result; generating a pre-interaction scheme based on the relevant information of the matched users, and analyzing the interaction data generated by the matched users in the interaction process to generate an interaction response scheme; and creating a virtual role based on the pre-interaction scheme, and generating a response of the virtual role in the interaction process based on the interaction response scheme so as to perform intelligent interview.

In the embodiment of the invention, an artificial intelligence mode is adopted, the expectation of a user is classified and calculated through a pre-established expectation analysis model to obtain an expectation analysis result, and the capability of the user is classified and evaluated through a pre-established capability level analysis model to obtain a capability analysis result; matching the user based on the expected analysis result and the ability analysis result; generating a pre-interaction scheme based on the relevant information of the matched users, and analyzing the interaction data generated by the matched users in the interaction process to generate an interaction response scheme; and creating a virtual role based on the pre-interaction scheme, and generating a response of the virtual role in the interaction process based on the interaction response scheme so as to perform intelligent interview, thereby solving the technical problem of low interview efficiency caused by manual interview.

Drawings

The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the invention without limiting the invention. In the drawings:

FIG. 1 is a schematic diagram of an intelligent interview system according to a first embodiment of the invention;

FIG. 2 is a schematic diagram of an intelligent interview system according to a second embodiment of the invention;

FIG. 3 is a schematic diagram of an intelligent interview system according to a third embodiment of the invention;

FIG. 4 is a schematic diagram of an intelligent interview system according to a fourth embodiment of the invention;

FIG. 5 is a schematic diagram of an intelligent interview system according to a fifth embodiment of the invention;

FIG. 6 is a schematic diagram of an intelligent interview system according to a sixth embodiment of the invention;

fig. 7 is a flow chart of a method of intelligent interviewing according to a seventh embodiment of the invention.

Detailed Description

In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are capable of operation in sequences other than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

Example 1

According to an embodiment of the present invention, there is provided an intelligent interview system, as shown in fig. 1, the system including: classification evaluation module 10, dispatch output module 12, interaction calculation module 14, and simulated interaction module 16.

1. Classification evaluation module

And the classification evaluation module 10 is configured to perform classification operation on the expectation of the user through a pre-established expectation analysis model to obtain an expected analysis result, and perform capability level classification evaluation on the capability of the user through a pre-established capability level analysis model to obtain a capability analysis result.

Wherein the desired analytical model is configured to be at least one of: analysis of job seeker: collecting historical interview data, and carrying out induction classification on successful and failed cases to extract contents concerned by job seekers; evaluating the extracted contents concerned by the user according to a multi-dimensional evaluation standard to generate different expectations of the job seeker; setting weights for different expectations, processing the different expectations based on the weights, and generating a final expected analysis result of the job seeker; analyzing recruiters: setting recruitment standards, and screening big data in three ranges of the same industry, the same post type and the same standard type according to the set recruitment standards to generate a scoring system; using the scoring system to perform weight scoring on the employment requirements of the recruiting positions according to different recruitment standards; analyzing the recruiting positions based on the weight scores to generate a final expected analysis result of the recruiter; wherein the recruitment criteria comprises at least one of: the recruiter sets standards, post transverse standards and industry transverse standards by the recruiter; wherein the user comprises the job seeker and/or the recruiter.

In one exemplary embodiment, the job seeker analysis model is further configured to: defining the type of the job seeker, marking the job seeker according to the type, extracting resume content of the job seeker, and acquiring a feature identifier; combining and sorting the feature identification and the big data to level the extracted features, and matching the extracted features with corresponding recruitment posts according to different levels; the fixed level comprises a first level direct demand, a second level direct related demand, a third level direct association demand, a fourth level indirect related demand and a fifth level indirect association demand.

In an exemplary embodiment, the capability level analysis model is configured to at least one of: classifying and analyzing competence levels of job seekers: based on the working ability, skill operating ability, communication ability, execution ability, logical thinking ability, character judgment and personality judgment of the job seeker, independently splitting and confirming the ability level of the job seeker, and carrying out measurement analysis on the split and confirmed ability level of the job seeker according to a plurality of dimensions of post standard, industry standard, environment standard, skill standard and experience standard to generate an ability evaluation analysis result of the job seeker; recruiter competency level classification and analysis: and generating a capability assessment analysis result of the recruiter based on the recruiting posts, the team, the work content, the directorywise, the enterprise situation, the promotion and the future development.

2. Dispatching output module

A dispatch output module 12 configured to match the user based on the desired analysis results and the capability analysis results.

Dispatch output module 12 is configured to interactively match, automatically match, answer online, and answer offline.

Interactive matching: based on the expected analysis result and the capability analysis result, sorting out data required by interaction, and simulating the interactive content through interaction matching output operation without real user participation;

automatic matching: based on at least one of: the method comprises the following steps of performing industry matching analysis, demand analysis, trend analysis, commonality analysis, general analysis and iterative analysis to generate classification and evaluation on the whole industry; and for at least one of: integrating the post content, the post advantages and the post characteristics, and matching and fitting the recruiting post and the job seeker again based on the classification and evaluation of the whole industry;

and (3) responding on line: generating an alternative interaction scheme through an instant interaction analysis model based on the interaction matching result and the automatic matching result, and performing alternative matching to perform online response;

off-line response: and performing automatic verification, automatic maintenance and automatic follow-up on the on-line response so as to maintain the output result of the whole on-line process.

3. Interactive computing module

And the interaction calculation module 14 is configured to generate a pre-interaction scheme based on the relevant information of the matched users, and analyze interaction data generated by the matched users in the interaction process to generate an interaction response scheme.

In an exemplary embodiment, the interaction computation module 14 is configured to at least one of: and (3) calculating the scheduled interaction: performing operation simulation of interactive contents on a user before an interactive process occurs based on a preset scheme, and outputting interactive contents for a user simulation interactive module to simulate an interactive situation to occur; and (3) instant interactive calculation: and performing operation analysis on data generated in real time in the interaction process, dynamically identifying user characteristics, and performing prejudgment on the basis of the identified user characteristics so as to output instant interaction content.

In an exemplary embodiment, the interaction computation module 14 is configured to at least one of: analog analysis of job seeker: generating a scheme for simulating the job seeker according to the interest analysis result, the ability level analysis result and the expected analysis result of the job seeker; performing side writing on the job seeker according to the scheme for simulating the job seeker, and determining the basic character and the psychological characteristic of the virtual role of the job seeker; performing simulation behavior analysis on the job seeker in the interaction process, and determining the behavior tendency of the job seeker; and (3) recruiter simulation analysis: after the post standard of the recruiter is determined, information is collected and integrated according to the conditions of the recruiter enterprises, teams and leader, and a recruiter simulation scheme is generated based on the integrated data.

4. Analog interaction module

And the simulation interaction module 16 is used for creating a virtual character based on the pre-interaction scheme and generating a response of the virtual character in the interaction process based on the interaction response scheme so as to carry out intelligent interview.

In an exemplary embodiment, the analog interaction module 16 further comprises an encoding module configured to: predicting a position of a triangular spacing line of a next block from a spacing line of a triangular wave of a previous block among adjacent data blocks of an image in an interactive process, such that the triangular spacing line at the predicted position forms an extension of the triangular wave of the adjacent data block to the next adjacent data block; determining an approximate direction of the extension region using randomly varying probabilistic binary entropy coding and a fixed bit length of a size of a last block of the adjacent data blocks; refining an approximate direction of the propagation region using the refinement information in the neighboring data blocks, wherein the approximate direction extends along a direction that fits a slope of the approximate direction of the propagation region by a previous one of the neighboring data blocks and has an offset that depends on an offset of the next one of the neighboring data blocks; encoding the image based on the refined approximate direction of the extension region.

In one exemplary embodiment, the encoding module is further configured to: determining a length of a prefix of a coded syntax element in the refinement information, and refining an approximate direction of the extension region based on an approximate direction of a maximum slope among a plurality of approximate directions of the prefix index and an approximate direction of an angle local maximum density among the plurality of approximate directions of the prefix index according to the determined length of the prefix; reconstructing a prediction error as an error correction signal based on the refined approximate direction of the extension region, and correcting a pre-coding matrix based on the prediction error to encode the image.

In this embodiment, an intelligent interview system using AI instead of manual interview is provided, which uses bilateral logic operations to perform simulated profile for job seekers and recruiters, respectively. The system establishes a desired analysis model, a capability level analysis model, a user interest analysis model, an instant interaction analysis model and a position depth analysis model. The job seeker and the recruiter are analyzed respectively, the fact that the user can simulate and learn according to the requirements of the user and the change of real-time information in the using process of the system is determined, unique experience and feeling are given to different users respectively according to different scenes, and the problem that the people cannot communicate in time due to time asynchronism frequently encountered in the modern society recruitment industry is solved.

Example 2

According to the embodiment of the invention, the intelligent interview system is provided, and a real user can be simulated by using a virtual role. Specifically, virtual characters are generated in the system for real users, and interviewing is completed by the virtual characters.

As shown in fig. 2, the intelligent interview system comprises: classification evaluation module 10, dispatch output module 12, interaction calculation module 14, simulated interaction module 16, and data block 18.

An expectation analysis model, a capability level analysis model, a job seeker interest analysis model, a recruiter interest analysis model, an instant interaction analysis model and a post depth analysis model are established in the intelligent interview system of the embodiment. And the analysis model is utilized to realize the multi-dimensional evaluation of the psychology, behavior, intention, demand, expectation and standard of the user, the user is subjected to side writing, and a virtual character closest to the actual situation of the user is shaped to realize the interaction capability in the interview process.

The classification evaluation module 10 is an initial module for analyzing the user, and needs to perform logical completion on only resume or post information according to limited information in combination with big data operation analysis, so as to play a role of a framework for building a virtual character.

The dispatch output module 12 fills the virtual character frame generated by the classification evaluation module 10 with contents, and makes the virtual character have the expectation and intention in accordance with the real user. And matching according to the generated expectation and the intention, and performing subsequent interview work.

The interaction calculation module 14 analyzes the whole interaction process of the job seeker and the recruiter by using role simulation analysis, and generates two parts of data, wherein one part of data is used for complementing the information of the user, and the other part of data is used for completing the interaction.

The simulation interaction module 16 is a place for how the generated virtual characters interact, wherein high requirements are required for interaction logic, response judgment and reaction time effectiveness of both interviewer parties.

Specific implementations of the various components are set forth in detail below.

1. Classification evaluation module

The classification evaluation module 10 is configured to perform desired demand classification, capability level classification, industry requirement matching, and special requirement matching.

1. The expected demand classification:

aiming at different role emphasis points of the user, expected analysis and classification evaluation are respectively carried out according to the role played by the user in the interview process.

(1) Job seeker desire requirement analysis and classification: firstly, defining the type of a job seeker, marking the job seeker according to the type, firstly, extracting resume contents of the job seeker, and obtaining a unique identifier from the resume contents. The unique identification and the big data are combined and arranged perfectly, and the first-level direct requirement, the second-level direct related requirement, the third-level direct association requirement, the fourth-level indirect related requirement and the fifth-level indirect association requirement are graded respectively. And matching the matched position information according to different level requirements.

(2) Analyzing and classifying the requirement of the recruiter: firstly, defining post employment standard, then screening big data in three ranges of same industry, same post type and same standard type according to the post standard. And comprehensively judging the types of the available personnel and grading, wherein the first-grade high matching degree, the second-grade medium matching degree and the third-grade low matching degree are high. Besides the first grading, the second and third grade type personnel need to add feature description, the feature of which aspect of the description content is the personnel can be used as an advantage to match the post standard, and the recruiter carries out final confirmation.

In this implementation, a neural network is used to build the expected analysis model.

The expected analytical model is: the vision of the user is analyzed and confirmed according to different requirements of the user in the place, and the vision is automatically updated by learning. The desire analysis model primarily analyzes the desires of the job seeker and the desires of the recruiter.

Job seeker desire analysis: the method comprises the steps of confirming the urgency degree of demand, the reason of demand, the demand standard, the demand elasticity, the degree of acceptable variation (risk), confirming the idea, the ability, the state, the target, the plan and the psychological expectation of a user, analyzing the evaluation of a third party to the user when possible, and finally combining different characteristics of the user to generate the expected analysis.

Recruiter desire analysis: and (4) setting post standards by the recruiter, and respectively performing weight setting on the timeliness, the capability matching degree, the soft skill requirement, the hard skill requirement and the personnel risk degree. And completing the weight items through user information input, bringing the weight items into an industry standard alternative library for large-scale comparison and performing mutual completion.

2. Classification of competency levels

The user information is collected in a multi-mode, the ability level of the user is evaluated according to the user information, and the evaluation content is different when the role of the user is different.

(1) Classifying and analyzing the competence level of the job seeker: the method comprises the steps of firstly accessing a job seeker, collecting relevant information of the job seeker and generating a horizontal competence report of the job seeker through a competence level analysis model. The emphasis of the system is to separate and confirm the working ability, skill and operation ability, communication ability, execution ability, logical thinking ability, character judgment and personality judgment of the job seeker. And carrying out double-aspect demand calling on the job seeker and the recruiter by combining the capability level analysis model to carry out comprehensive evaluation.

(2) Recruiter competency level classification and analysis: firstly, enterprise information is collected, and the situation of the post is analyzed according to big data information. Respectively determining posts, teams, work content, direct leaders, enterprise conditions, promotion, future development and the like. The welfare contents provided by posts, teams, employees and the like in an enterprise are comprehensively evaluated, the preference and the interest of the posts are determined through a capability level analysis model, and the comprehensive evaluation is carried out by combining the demand call of the job seeker and the recruiter.

In this embodiment, a competency level analysis model is built and trained using neural networks to analyze the competency levels of recruiters and job seekers.

Capacity level analysis model: and determining the ability level of the user according to the information provided by the user and the communication condition of the user. And (4) calling the post standard and combining with the comparison and analysis of the industrial post, determining the basic information of the user such as skill, experience, academic history, psychology, communication, personality and the like, simultaneously investigating the previous work of the user, simulating and confirming the real situation of the user, and finally outputting the comprehensive evaluation result of the user.

3. Matching industry requirements: information collection is carried out aiming at the posts of the same type in the industry, and matching simulation is carried out through the information collected in the interview process of a plurality of posts of the same type. And performing matching simulation on the information of both the job seeker and the recruiter output by the expectation analysis model and the capability level analysis model.

4. And (3) special matching classification: because some posts or job seekers have special requirements, they need to be specially defined according to different requirements, and determine advantages, disadvantages and contents suitable for pursuing. For example, since job seekers like a busy work environment, the industry associated with the demand may be considered at this time for recommendation matching.

2. Dispatch output module 12

Dispatch output module 12 is configured to implement an interactive match output, an automatic match output, an online answer, and an offline answer.

1. And (3) interactive matching output: and combining the output results of the expected analysis model and the capability level analysis model to carry out simulation interaction on the possible intentions of the two parties, wherein the interaction belongs to pure simulation, real users are not required to participate, the details of the interaction result are not actively displayed for the users, and auxiliary display is carried out only when one party of the users formally participates in the interaction. After the formal interaction is finished, the result of the interaction is combined to carry out contrastive analysis, so that both parties have clear understanding to confirm the situation of the parties.

2. And (3) automatic matching output: and collecting the total data of both the two parties of the interview, then performing operations such as sorting, cleaning, analyzing, judging, classifying and the like, and then performing matching fit again on the user through calculation, wherein the matching fit comprises post matching, experience matching, expectation matching, capability matching and requirement matching, gathering to generate a complete matching relationship recommendation mechanism and outputting a matching result.

3. And (3) on-line response output: and judging the content possibly appearing in the online interaction process through an instant interaction analysis model, and performing dynamic expectation judgment, dynamic expectation range judgment, special expectation judgment and recommendation acceptance judgment. And then carrying out matching association on the expectation, wherein the matching association comprises dynamic expectation association, dynamic expectation matching, dynamic expectation range judgment, special expectation association, special expectation matching and recommendation acceptance record.

4. And (3) offline response output: collecting information in the traditional interview process comprises personnel verification and post verification, and maintaining and automatically following related information after verification. And generating information required for perfection, perfecting the online information and simultaneously inputting the online information into a database to become data content of the big data service.

In the embodiment, a neural network is adopted to establish a job seeker interest analysis model and a recruiter interest analysis model.

The interest analysis model of the job seeker comprises: according to the result obtained by the expectation analysis, the intention of the job seeker is determined, and then the interest of the user is determined according to the feedback of the simulated interview and the communication before the interview. For example, the post where the user interviews is an attendant, but the interest is reading and writing, at which time posts like book bars, cafes, etc. would be more appropriate.

Recruiter interest analysis model: and determining the interest preference of the recruiter by taking the standard obtained by the expected analysis model as a premise and combining the communication content before interviewing. For example: the job team leader likes the staff with neat image, and even if the conditions of various aspects of job seekers meet the standard, the different definitions of the cleanliness need to be confirmed through analysis.

3. Interactive computing module

The interactive computation module 14 implements predetermined interactive computation and instant interactive computation.

1. Predetermined interaction computation

And performing preset content design on the upcoming interactive content according to the information collected from multiple aspects, and performing simulation analysis on the job seeker and the recruiter after an interactive scheme is output.

(1) Analog analysis of job seeker: and analyzing the behavior tendency of the job seeker by combining the results of the expected demand analysis, the capability level analysis and the comparison and analysis of the same type of personnel. After the analysis is completed, the job seeker is subjected to side writing firstly, the setting of the role is determined, then the role of the job seeker is simulated, the quality of the job seeker required to be subjected to interview is simulated, the advantage performance of the job seeker is improved, and the disadvantage performance caused by tension is eliminated.

(2) And (3) recruiter simulation analysis: and setting the role of the recruiter according to the set standard and the evaluation of the post comprehensive information obtained after data collection and analysis, and simulating a virtual role which is easily accepted by people. And after the virtual role is created, the bottom layer requirements of the recruiter are deeply mined, the contents needing to be known in the interviewing process are perfected, and the standards are dynamically updated and upgraded, so that the purpose of perfecting the role is achieved.

2. Instant interactive computing

In the interview interaction process of the user, different judgment and control are carried out on the interactive content according to different positions of the virtual character, so that the interaction purpose is achieved, and meanwhile, the spread of the interactive content is avoided. Extension association is allowed, but extension of a taste will lead to the spreading of the content, requiring an extension boundary to be set for definition. In addition, the behavior intention of the user is judged and analyzed in the interaction process, and the idea of the user is defined through an analysis model. Determining the requirement of a user, judging the intention of the user after judging the requirement of the user, confirming the current idea, state, goal and plan of the user, comprehensively evaluating the acceptance degree and the possibility of the intention of the user to the post, and finally outputting a response scheme required by the virtual character in the interaction process.

In this embodiment, a position depth analysis model is built and trained using a neural network. According to the definition of the position, the position is deeply analyzed, including but not limited to industry development trend, human gap trend, capability standard curve and the like. And after data expansion and user input requirements, performing analysis content completion through learning to generate complete post SWOT analysis. Meanwhile, the bottom layer requirements of the personnel units are deeply excavated, so that the character simulation can be closer to reality.

4. Analog interaction module

The simulation interaction module 16 primarily implements recruiter simulation and job seeker simulation.

Recruiter simulation: and creating a recruiter virtual role, establishing a role frame according to the previous module output, and setting the requirement by a real user. And collecting the post information analysis result, and determining the information matched with the post through the post depth analysis model. And calling a preset interaction output result to carry out interaction logic filling on the virtual character, and finally generating the virtual character for interviewing. The role needs to be capable of deeply analyzing conditions of job seekers while meeting communication, generating an analysis report and being used for selection analysis of a final user.

Job seeker simulation: and creating a virtual role of the job seeker, establishing a role framework according to the previous module output, and setting requirements by a real user. According to the setting of user requirements, defining the user, extracting user characteristics to start for transverse comparison operation, determining the advantages and disadvantages of the user, and then performing longitudinal comparison according to the user intention to determine the opportunity and threat of the user. Firstly, a report is generated to the user for self-identification, for example, the user completes the creation after confirming to be feasible and logically fills the virtual role with the generated result, and finally, the virtual role which can represent the user of the job seeker is generated.

5. Database with a plurality of databases

The database 18 is primarily configured to store various data of the user, including various data generated at the time of the intelligent interview.

The virtual role application system based on artificial intelligence in the field of human resource recruitment provided by the embodiment of the invention can replace manual interview, complete the interview process through the virtual role when personnel cannot interview, and feed back the result.

Example 3

According to the embodiment of the invention, the intelligent interview system is provided, and the intelligent interview system can simulate roles of two parties in a recruitment process to realize simulated instant communication. As shown in fig. 3, the intelligent interview system comprises a classification evaluation module 10, a dispatch output module 12, an interaction calculation module 14 and a simulation interaction module 16.

1. Classification evaluation module

The classification evaluation module 10 is used for performing classification evaluation on information already existing in the system and performing evaluation refinement according to the completion information generated after the consultant communicates. Respectively performing side writing on job seekers and recruiters, performing simulation operation on expected requirements of users, performing classification operation through an expected analysis model to generate a simulated role form of the users, generating a corresponding question bank for standby, paying attention to the problem of the users who pay attention to the operation result, and performing special marking.

And after the simulation is finished, classifying and evaluating the capability level of the user, generating the existing capability evaluation report of the user through a capability level analysis model, and discovering capability evaluation results such as a potential evaluation report. And after the capability evaluation result is generated, carrying out industry requirement matching aiming at the capability of the user, not only aiming at a certain post, but also simulating an industry classification condition through data operation, generating a matching logic according to the industry requirement, and combining the generated user evaluation result and side writing simulation to completely generate a role simulation of the user.

In the classification evaluation module 10, two models are mainly included: a desired analytical model 102 and a capability level analytical model 104.

It is desirable that the analytical model 102 be able to correctly recognize the expectations of the user. The past data is collected first, and induction and classification are carried out on successful and failed cases to form input. Extracting the content concerned by the user, judging according to multi-dimensional evaluation standards such as high frequency, important attention, multiple mentions, decisive factors and the like, and generating and outputting. And then, weight setting is carried out on the expected content, weight combination is carried out on different crowds and industries according to the big data analysis result, and finally an analysis result is generated.

The capability level analysis model 104 may generate capability level criteria for the post demand. And aiming at the situation that a job seeker needs to collect user data in a large range, and the job seeker is subjected to measurement analysis according to multiple dimensions of post standards, industry standards, environment standards, skill standards and experience standards. The related position expectation is extended according to the expected position of the job seeker, then the industry expectation is obtained, and finally the output is the capability evaluation analysis summary of the job seeker.

2. Dispatching output module

The dispatch output module 12 is used for dispatching the next things to be done by the user-generated character. Dispatch output module 12 is configured to implement four functions of interactive matching, automatic matching, on-line answering, and off-line answering.

Interaction matching output requires a user to expect an analysis model and an ability level analysis model to sort out data required by interaction, possible interaction contents are simulated through interaction matching output operation, a simulation scheme is stored, and in each interaction process, a real character is replaced and borrowed as a virtual character and interacts with the virtual character.

The content output by the automatic matching calculation is data matching obtained in the whole background database, wherein the data matching comprises industry matching analysis, demand analysis, trend analysis, commonality analysis, general analysis, iterative analysis and the like, and the generated result is classification and evaluation of the whole industry.

The post matching analysis needs to integrate the feature contents such as post content, post advantages, post characteristics and the like, and then match the post with job seekers by combining with industry analysis. The final content output by the system is a recruiter interest analysis model.

There is a need for a match analysis of job seekers on the recruiter side, where the primary analysis is the availability of job seekers. The most problem faced by the recruitment industry at present is that the availability cannot be correctly acquired during the interviewing process. Therefore, manpower is wasted, the job seeker needs to be subjected to side writing through the current job seeker information, and finally, a job seeker interest analysis model is output. And after the two matching results are generated and output, respectively distributing and outputting the results to an online response operation module and an offline response operation module according to the situation requirements.

The online response module needs to generate an alternative interaction scheme through the timely interaction analysis model and perform alternative matching.

The offline response module needs to be able to perform automatic verification, automatic maintenance, automatic follow-up, and the like to maintain the output result of the entire online process.

In this embodiment, whether the job seeker interest analysis model can correctly perform simulation analysis on job hunting preference and interest of the job seeker is determined. The job seeker is first defined to determine what type the job seeker belongs to. For example: aggressive type, adventure type, mucous type, depressive type, etc. After definition, the job seeker is subjected to type analysis, such as: the radical job seeker prefers, pays attention to, has advantages and capability to play and the like. The analysis of the genre is followed by a determination and interpretation of the interests of the job seeker, for example: the enthusiast caregivers read the working atmosphere of the enthusiasm determined by the interest judgment into colleagues with enthusiasm in the working environment, and do not pay attention to whether the leaders have the enthusiasm or not. Or a leadership with enthusiasm, and whether the colleagues are enthusiasm or not is not concerned. And the model learning and perfection are carried out by analogy.

In this embodiment, the recruiter interest analysis model needs to accurately define the actual skill needs of the personnel for the employee, and then extend other contents on the premise of meeting the capability. For example: the staff of the trainee station and the colleagues who need the physical health of the staff have good classification and arrangement capability. After the classification and arrangement capabilities are met, the requirements of good communication of staff, good team cooperation spirit and the like are extended. After the requirement of work is confirmed, environmental factors of work, growth factors of posts, planning factors of future development of personnel and the like are combined to generate a result which is attractive to job seekers and can meet the future planning of the posts and the requirement of employment in a long term or a short term.

3. Interactive computing module

The interaction computation module 14 is a part that performs data operations when the simulated character interacts with the user. The system comprises a preset interactive calculation module and an instant interactive calculation module.

The preset interaction calculation module is mainly used for preprocessing before an interaction process occurs, generating a role simulation and performing side writing on an interactive real object to simulate a possible interaction situation. And performing operation simulation on interactive contents of the user in a frame of a set preset scheme through job seeker simulation and recruiter simulation, and finally outputting the interactive contents which can be used for a user simulation interactive module.

The instant interactive computation module is used for carrying out data operation processing when an interactive process occurs, wherein the data operation processing comprises two types of text instant messaging and multimedia instant messaging. And carrying out operation analysis on the data generated in real time in the interaction process, and carrying out prejudgment. And on the premise of realizing the interaction, acquiring a final judgment result of whether the actual conditions and the requirements of the two interactive parties are in agreement or not. And recommends associations that may exist with the data in the process.

The real-time interactive calculation needs to be combined with the existing AI communication function, output results by matching with analysis models of job seekers and recruiters, and support ongoing interaction, such as requirement judgment, requirement mining and the like. The user characteristics are dynamically identified in the interaction process, and prejudgment is carried out on behavior intention changes which may occur to the user in the interaction process, such as: new skills, new experiences, new ideas, new requirements. And the requirement matching is carried out according to the dynamic change of the user condition. And collecting fragment information generated in the interaction process for integration, combining the fragment information with original information of the user according to the integrated information, and generating new user information for defining the user after combination.

In this embodiment, the instant interaction analysis model needs to be able to analyze the problems that may be encountered during the instant interaction process and solve the problems in advance. For example: the job seeker has no accommodation, and the job seeker has the accommodation requirement, so that a scheme for solving the accommodation problem needs to be prepared in advance to deal with the problem faced by the job seeker. The output content of the instant interaction analysis model is a problem solution which is possibly generated before and during the interaction process.

4. Analog interaction module

The simulation interaction module 16 is a module which is most intuitively experienced by users in the whole system, wherein recruitment simulation is started according to the requirement of a job seeker, a simulation role is generated through industry requirement setting, output of a recruitment interest analysis model, output of an instant interaction analysis model, output of an expectation analysis model and output of recruitment simulation analysis to serve as an interviewer, and the candidate meeting the job position is subjected to functional interview.

A post depth analysis model needs to be generated during interviewing, data content is perfected as much as possible in the process of repeated interviewing, and interaction with subsequent candidates is better performed through learning simulation. In the face of recruiters, the simulation of job seekers needs to be started, a virtual role of the job seeker is generated through industry requirement setting, output of an expected analysis model, output of a capability level analysis model, output of a job seeker interest analysis model, output of an instant interaction analysis model and output of simulation analysis of the job seeker, interviews an enterprise with a recruitment requirement, and positioning is generated by combining communication with the job seeker, so that the determination of the real situation of the job seeker by the enterprise is met.

In this embodiment, the job seeker simulation function may perform side-writing on the job seeker according to the output result of the analysis model, generate a virtual character, and simulate the job seeker to perform interview communication with the recruiter. The side writing result is required to be as accurate as possible, and the real will of the job seeker can be expressed. The main purpose is to enable the job seeker to complete the interview process of the job seeker through the virtual role at any time.

And the recruiter simulation function is to perform side writing on the recruiter according to the analysis model output result of the recruiter, generate a virtual role and simulate the interview communication between the recruiter and the job seeker. The result of the side writing needs to reflect the real expectation of the recruiter for the job application standard of the recruit post and express the real expectation for the special expectation of the employee, such as the character and the like. And in the recruitment process, comprehensively analyzing the capability of the candidate, and comparing by combining the preset weight score. And generating result output which can be delivered to the enterprise while realizing the function of simulating the communication between the character and the candidate. And evaluating the result of the interview by the recruiter, and subsequently perfecting the unrecited content.

The virtual role application system is created in the field of human resource recruitment based on artificial intelligence, interviewing can be performed instead of manual work, interviewing is completed through the virtual roles when interviewing cannot be performed by personnel, and results are fed back. Thereby improving the efficiency of the interview.

Example 4

According to the embodiment of the invention, the intelligent interview system is provided, and the customized virtual role is generated in the system by using an AI simulation technology to conduct interview on behalf of a user. And through the parallel mode of multichannel, carry out a plurality of simulation interviews simultaneously to the interview efficiency has been promoted greatly. In addition, the system can be on-line standby for 24 hours, so that the problem that the interview time and the interview place cannot be matched at the initial stage of the interview can be well solved.

As shown in fig. 4, the intelligent interview system comprises a classification evaluation module 10, an interaction calculation module 14, a dispatch output module 12 and a simulation interaction module 16, wherein the simulation interaction module 16 comprises a coding module 162.

1. Classification evaluation module

The triage evaluation module 10 is configured to perform a desired need analysis for recruiters and job seekers and to perform an analysis of the competency level of the job seekers.

Analysis of the expected demand: the expectation requirement analysis is integrally divided into two parts, namely, the analysis of the expectation requirement of the job seeker and the analysis of the expectation requirement of the recruiter.

The job seeker analysis first requires definition of the job seeker, determining how strongly the job seeker's will and what the vision for new work is. And marking the job seeker according to the classification screened out by the big data, setting the weight according to different expected contents of the job seeker, and generating detailed descriptions of the job seeker through an expected analysis model so as to enable the system to carry out the next operation.

The recruiter analysis requires setting standards including self-setting standards, post horizontal standards, and industry horizontal standards. After setting the standard, defining the recruited post according to the standard content, wherein the defined content is to generate a scoring system, perform weight scoring on the job occupation demand according to different standard contents, generate the demand condition of the post through an expected analysis model, and deliver the demand condition to the system for the next operation.

Analysis of the competence level: after the post demand is generated, comprehensive assessment of the job seeker's ability level is initiated. Firstly, whether the job seeker is qualified for the post or not is confirmed, and a single work skill cannot be used as the total weight of the project. The ability of the job seeker needs to be determined based on the characteristics of the post. Such as on duty cycle, on duty stability. The mental states of the people include characters, communication, potential and the like. And comprehensively evaluating the job seeker through the capability level analysis model and combining the output result of the job seeker part in the expected analysis model to generate more comprehensive data for the next operation of the system.

The classification evaluation module 10 can be configured to analyze the interests of the job seeker and the recruiter in addition to analyzing the desires of the job seeker and the recruiter.

Analyzing interest of job seekers: the interest analysis model of the job seeker generates post content matched with the job seeker according to the comprehensive evaluation result of the expected combining ability of the job seeker, and performs associated matching recommendation. According to different states of the job hunters, different communication modes need to be generated for different job hunters in the simulation communication process, so that the communication needs are met. During the period that the job seeker exists in the system, all interview data can be reserved and analyzed, and the situation of the job seeker is corrected according to the situation of each interview, so that the purpose that the real situation of the job seeker is most close to is achieved.

Analyzing the interest of the recruiter: the recruiter interest analysis model is an analysis of the needs of the recruiter in addition to the position criteria. The method comprises the expectation of a team belonging to the post on the personnel, the expectation of a department on the post, the soft requirement of a direct leader on the personnel and the like. Information collection is firstly carried out according to the expectations and the requirements, then interview feedback of each time is analyzed in the process of simulating interview, and the actual requirements of the recruiter and the post are combined and corrected. The final purpose is to allow the system to correctly analyze the bottom-level requirements of the recruiters and find suitable job seekers according to the requirements.

2. Interactive computing module 14

The interaction calculation module 14 is configured to perform computation analysis on data generated in real time during an interaction process through real-time calculation, dynamically identify user characteristics, and perform prejudgment based on the identified user characteristics.

One of the primary functions of the interaction computation module 14 is instant interaction analysis.

And (3) instant interactive analysis: and performing operation analysis processing on all resources required in the interaction process, analyzing the psychological expectation of the user, feeding back according to an expected result, realizing the guidance expected by the user, and guiding the psychological expectation of the user to the direction with higher matching degree. And analyzing the current state of the user, and sorting background information according to the current state to perform two operations, wherein one operation is to correct the user information, and the other operation is to adjust a response scheme according to the current state of the user. And analyzing and integrating the static information of the user, the user behavior record and the information generated by the ongoing interaction, and then matching to obtain the optimal matching degree, and completing the role simulation function by the output content.

3. Dispatching output module

Dispatch output module 12 is configured to perform interaction scheme matching.

And (3) interaction scheme matching: the interactive scheme is to sort and classify the collected data, define and extract the user characteristics, respectively analyze and judge, and call corresponding modules according to the analysis result. And generating an interaction scheme through operation analysis after the analysis judgment, and correcting the interaction scheme according to the analysis result of the data information generated by the user in the interaction process.

4. Analog interaction module

The simulation interaction module 16 is configured to simulate the job seeker and the recruiter, and basically comprises: analog analysis of job seekers, analog analysis of recruiters, role simulation of job seekers and role simulation of recruiters.

Analog analysis of job seeker: and generating a simulation scheme after integrating according to the interest analysis result, the ability level analysis result and the expected analysis result of the job seeker. And the job seeker is subjected to side writing according to the simulation scheme, the basic character and psychological characteristics of the virtual character of the job seeker are shaped, the simulation behavior analysis is carried out on the job seeker in the communication process, the behavior tendency which may occur to the job seeker is determined, and a corresponding solution is designed. And correcting the contents of the simulation scheme in the subsequent job seeker information completion process.

And (3) recruiter simulation analysis: and after the post standard of the recruiter is determined, information is collected and integrated according to the conditions of the recruiter enterprises, teams and leader. And generating a simulation scheme by the integrated data, and modifying and perfecting by the recruiter.

Role simulation of job seeker: and performing virtual character creation on the job seeker by combining an interaction scheme preset in advance, and performing final confirmation by the job seeker after the creation is completed to realize that the virtual character can represent the intention of the job seeker for communication. The content created by the character includes voice, ability description, response skill, character simulation, context simulation, emotion simulation, experience cognition, expression logic, response logic, and the like. The definition of the simulation of the role of the job seeker is to create a virtual role in the system to completely represent the job seeker for interviews.

And (3) recruiter role simulation: and (4) creating the role of the recruiting interviewer according to a preset simulation scheme, wherein in the creating process, the recruiter firstly sets the requirement and carries out deep analysis by combining the post responsibility requirement. Including ability assessment criteria for job seekers and deep job requirements of enterprises. The analysis result is used for generating a virtual role, defining types for the role, and attracting job seekers to conduct interviews according to the value of the enterprise and the characteristics of recruiting the enterprise.

In the present embodiment, the analog interaction module 16 further includes an encoding module 162, and the specific operation of the encoding module 162 is as follows.

Predicting a position of a triangular spacing line of a next block from a spacing line of a triangular wave of a previous block among adjacent data blocks of an image in an interactive process, such that the triangular spacing line at the predicted position forms an extension of the triangular wave of the adjacent data block to the next adjacent data block; determining an approximate direction of the extension region using randomly varying probabilistic binary entropy coding and a fixed bit length of a size of a last block of the adjacent data blocks; refining an approximate direction of the propagation region using refinement information in the neighboring data blocks, wherein the approximate direction extends along a direction that fits a slope of an approximate direction of the propagation region by a previous one of the neighboring data blocks and has an offset that depends on an offset of a next one of the neighboring data blocks.

The approximate direction of refinement of said extension region using refinement information in said neighboring data blocks is mainly made use of coded syntax elements in the refinement information. For example, a length of a prefix of the coded syntax element in the refinement information is determined, and according to the determined length of the prefix, an approximate direction of the extension region is refined based on an approximate direction of a maximum slope among a plurality of approximate directions of the prefix index and an approximate direction of an angular local maximum density among the plurality of approximate directions of the prefix index.

After the approximate direction of the extension region is refined, a prediction error as an error correction signal is reconstructed based on the refined approximate direction of the extension region, and a pre-coding matrix is corrected based on the prediction error to encode the image.

By the encoding method, the approximate direction of the extension area of the data block in the image can be found, so that the image encoding can be carried out based on the accurate approximate direction. Therefore, when the image is decoded and displayed by simulating the interview, the image can be displayed more clearly, so that the method is beneficial to observing the fine expression or action of the job seeker or the recruiter in the interview process, and the interview effect is improved better.

Through the structure, the embodiment of the application realizes the following beneficial effects:

1. and a virtual role is established in the system, so that the virtual role is used for replacing any one of the two parties when the interview cannot be performed in the interview process.

2. The interview efficiency is improved, a role can conduct multiple interviews simultaneously, and a detailed summary report is generated to assist a user in adjudicating interview results.

3. The matching and searching efficiency of the recruitment interview is improved, the information which is expected to be matched with the recruiting interview is searched in the database more quickly, and the next operation is carried out.

4. The efficiency of information data utilization is improved, and interview operation can be performed while information is generated each time. The efficiency experience of the user is improved, so that the user can obtain useful data information at the first time.

5. Deepens the communication depth of the interviews and generates a detailed content analysis report, so that the two interviews can better know the advantages of the two interviews and the information content related to the interviews, and the two interviews can be more clearly and definitely known.

6. Errors possibly caused by human factors in the interview are abandoned, manual work is replaced by artificial intelligence, and few errors or even zero errors are achieved.

Example 5

Fig. 5 is a schematic structural diagram of a deep interactive AI intelligent interview system according to a fifth embodiment of the invention. The structure of the intelligent interview system in this embodiment may be similar to that of any of embodiments 1 to 4, except that the intelligent interview system in this embodiment utilizes interdependencies between voice data and response data to determine the desires of the user.

As shown in fig. 5, the expected analysis model of the intelligent interview system includes a speech analysis processing module 52 and an auto-answer analysis processing module 54, wherein the speech analysis processing module 52 includes a dependency determination module 524, and the auto-answer analysis processing module 54 includes an expected analysis module 542.

The dependency determination module 524 reads a syntax structure having an interdependence relationship from the data stream of the voice data and the response data, wherein the syntax structure represents an interdependence between at least one pair of different values among the voice data and the response data; and based on the syntactic structure, using a truncated unary code to predict on different layers of different levels of information amount of statement elements corresponding to the voice data and the response data by using a structure, secondarily encoding the voice data and the response data in the interaction process, and normalizing the secondarily encoded data, wherein each statement element of each piece of voice data and each piece of response data is associated with a corresponding one of the different layers, and the different layers include a base layer, an extension layer and an additional layer. Wherein the structure prediction is a preset rule for predicting the interdependency by using a grammar structure.

The adopted interdependencies may be calculated based on the following manner.

In the above formula, x _i Is the information amount, y, of the speech data sentence element _i Is the information quantity of the statement elements in the response data, n is the total number of the statement elements in the voice data and the response data,

is the average of n sentence elements, a represents the learning rate, and b represents the interdependence.

The dependency determining module 524 performs secondary encoding on the speech data and the response data in the interaction process, and performs normalization processing on the data after the secondary encoding, on different layers of different levels of information amount of the sentence elements corresponding to the speech data and the response data, using the structure prediction, based on the calculated interdependency, using the truncated unary code. In the secondary encoding, each sentence element of each piece of voice data and each piece of answer data is associated with a corresponding one of the different layers.

By adopting the processing mode, statement elements with different interdependencies can be corresponding to different layers. So that the intention of the user can be analyzed intelligently through the neural network model at a later stage.

The neural network model provided in the present embodiment will be described in detail below.

The input layer of the neural network topology structure adopted by the embodiment of the invention is provided with three neurons, namely the information quantity of statement elements of voice data, the information quantity of statement elements of response data and interdependence. The hidden layer has six neurons, and the output layer has four neurons.

Setting the syntax element of the voice data as x and the syntax element of the response data as y, the network input layer to the output layer can be expressed as:

wherein, b ₁ For input layer to hidden layer threshold vectors, b ₂ For the hidden-to-output threshold vector, f (-) is the nonlinear function employed by the hidden layer, and g (-) is the output layer activation function. w is the weight matrix from the network input layer to the hidden layer, v is the weight matrix from the hidden layer to the output layer, z represents the output vector, P _mx Is a corresponding layer corresponding to the statement elements, upsilon is an influence factor, n is the total number of the statement elements in the response data and the voice data, F _best For optimal interdependence, i.e. the highest degree of interdependence, F _avgn The average of the interdependencies.

The dependency determination module 524 performs AI intelligent analysis on the normalized data using the neural network model described above. The expectation analysis module 542 generates response data in response to the voice data based on the intelligent analysis result.

In the embodiment, by introducing the interdependency between the syntax elements and the neural network model, the intention or the desire of the user can be analyzed more intelligently.

Example 6

Fig. 6 is a schematic structural diagram of an intelligent interview system according to an embodiment of the invention. The structure of the intelligent interview system in this embodiment may be similar to that of any of embodiments 1 to 4, except that the intelligent interview system in this embodiment can analyze the intention and desire of the user based on the voice intensity of the user.

As shown in fig. 6, the expected analysis model of the intelligent interview system includes a speech analysis processing module 52 and an auto-answer analysis processing module 54, wherein the speech analysis processing module 52 includes an intensity determination module 522, and the auto-answer analysis processing module 54 includes an expected analysis module 542.

The intensity determining module 522 obtains the maximum value, the minimum value and the average value of the voice intensity in the voice data in the previous voice window of the voice data; based on the maximum value, the minimum value and the average value of the voice intensity in the voice data in the previous voice window, normalization processing is carried out on the voice intensity data in the current voice window, whether a sampling point corresponding to the voice intensity data after normalization processing is a zero point or not is judged, and therefore two adjacent zero points in the current voice window are found out; calculating the voice intensity of the voice data in two moments corresponding to the two adjacent zero points based on the two adjacent zero points; and based on the voice intensity, carrying out slicing processing on the voice data in the current window, and setting a next voice window.

Wherein, the following formula can be adopted to carry out normalization processing on the data in the current voice window range:

in the formula: h _i The data values are normalized; h is _i The original value of the voice intensity of the voice data in the current window range is obtained; h is _i-1ave 、h _i-1max And h _i-1min The average value, the maximum value and the minimum value of the voice intensity data in the range of the previous window are respectively.

The normalized waveform of the voice intensity is similar to a sinusoidal image, and therefore, the feature values near the zero-crossing point are more easily distinguished than the feature values of the voice intensity near the peak and the trough. Therefore, the value near the extraction zero point is selected as the characteristic value of data analysis, so that the error caused by sampling is reduced, and the voice intensity is obtained more accurately. And multiplying the voice intensity values of the adjacent sampling points on the left side and the right side, and if the obtained result is a negative value, judging that the point is a zero point. The period of the sine wave of the voice intensity can be obtained through the images in the voice window passing through two adjacent zero points, and then the voice intensity can be calculated based on the period of the sine wave. Meanwhile, the situation that data waveform jitter of voice intensity occurs near the zero point and abnormal data occurs is considered, so that the abnormal data can be eliminated. For example, when the distance between two zeros is too close, which causes the calculation to be biased, the set of zeros is deleted to delete the abnormal value.

Since the speech intensity will change gradually, the time range of the next window can be set to be an integral multiple of the single period time of the last group of sinusoidal images in the current window, so as to ensure that the time range of the next window at least includes one sinusoidal period. Thus, the voice intensity of each window is calculated one by one until all voice data are calculated.

In this embodiment, the time range of the speech in the current window is set by the last group of sinusoidal cycles calculated by the previous window, then the data in the current window is normalized by the maximum value, the minimum value and the average value of the data in the previous window in the set time range of the current window, the speech intensity at each moment in the current window is calculated, the time range of the next window in the current window is set by the last group of cycles in the current window, and so on until all the speech data are calculated, so as to accurately obtain the speech intensity in real time.

According to the method, the voice intensity can be accurately calculated, and the time setting range of the next window can be set based on the voice intensity of the previous window, so that the window of the voice data can be divided more accurately, and the voice intensity can be calculated more accurately.

The expectation analysis module 542 calculates an expectation of the user based on the calculated speech intensity. The expectation analysis module 542 calculates the expectations of the user based on several dimensional vectors, such as arousal and alertness. Wherein, the arousal degree represents the height of the arousal degree, and the positive degree represents the height of the positive emotion. Both dimensions represent his height by a numerical value.

In one exemplary embodiment, the arousal may be calculated by the following formula:

wherein mp _{q table} Showing the arousal degree between two adjacent zero points, p showing the moment of the first zero point in a pair of zero points, and q showing the moment of the second zero pointX is the speech intensity, y is the speech rate, I represents the correction factor, and B represents the speech data.

In one exemplary embodiment, the aggressiveness may be calculated by the following formula:

both dimensions represent his height by a numerical value. For example, a numerical range of [ -1,1], with closer-1 representing a greater degree of fan/passive and closer-1 representing a greater degree of activation/activity.

After calculating the arousal and aggressiveness, the expectation analysis module 542 further extracts a speech element of the speech data between the two zeros and takes the speech element as an expectation feature value to determine the expectation of the user.

In the embodiment, the setting range of the next time window is set based on the previous time window, so that the voice data can be sliced more accurately, the voice intensity can be calculated, and then, the expected characteristic value can be extracted accurately by introducing two measurement parameters of the arousal degree and the positive degree, so that the expectation of the user can be analyzed accurately.

Example 7

According to an embodiment of the present invention, there is also provided an intelligent interview method, as shown in fig. 7, the method including:

step S702, classifying and calculating the expectation of the user through a pre-established expectation analysis model to obtain an expectation analysis result, and performing capability level classification and evaluation on the capability of the user through a pre-established capability level analysis model to obtain a capability analysis result;

step S704, matching the user based on the expected analysis result and the capability analysis result; generating a pre-interaction scheme based on the relevant information of the matched users, and analyzing the interaction data generated by the matched users in the interaction process to generate an interaction response scheme;

step S706, based on the pre-interaction scheme, creating a virtual character, and based on the interaction response scheme, generating a response of the virtual character in the interaction process, so as to perform intelligent interview.

The method in this embodiment can implement the functions of the systems in embodiments 1 to 6, and is not described here again.

The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.

Claims

1. An intelligent interview system, comprising:

the classification evaluation module is configured to perform classification operation on the expectation of the user through a pre-established expectation analysis model to obtain an expected analysis result, and perform capability level classification evaluation on the capability of the user through a pre-established capability level analysis model to obtain a capability analysis result;

a dispatch output module configured to match the user based on the desired analysis results and the capability analysis results;

the interaction calculation module is configured to generate a pre-interaction scheme based on the relevant information of the matched users, and analyze interaction data generated by the matched users in the interaction process to generate an interaction response scheme;

the simulation interaction module is configured to create a virtual character based on the pre-interaction scheme and generate a response of the virtual character in the interaction process based on the interaction response scheme so as to carry out intelligent interview;

the analog interaction module further comprises an encoding module configured to:

predicting a position of a triangle spaced line of a next block from a spaced line of a triangle wave of a previous block among adjacent data blocks of an image in an interactive process such that the triangle spaced line at the predicted position forms an extension of the triangle wave of the adjacent data block to the next adjacent data block;

determining an approximate direction of the extent using randomly varying probabilistic binary entropy coding and a fixed bit length of a size of a last one of the neighboring data blocks;

refining an approximate direction of the propagation region using refinement information in the neighboring data blocks, wherein the approximate direction extends along a direction that fits a slope of an approximate direction of the propagation region by a previous one of the neighboring data blocks and has an offset that depends on an offset of a next one of the neighboring data blocks;

encoding the image based on the refined approximate direction of the extension region;

determining a length of a prefix of a coded syntax element in the refinement information, refining an approximation direction of the extension region based on an approximation direction of a maximum slope among a plurality of approximation directions of an index of the prefix and an approximation direction of an angle local maximum density among the plurality of approximation directions of the index of the prefix according to the determined length of the prefix;

reconstructing a prediction error as an error correction signal based on the refined approximate direction of the extension region, and correcting a pre-coding matrix based on the prediction error to encode the image.

2. The system of claim 1, wherein the desired analysis model is configured as at least one of:

analysis of job seeker: collecting historical interview data, and carrying out induction classification on successful and failed cases to extract content concerned by job seekers; evaluating the extracted contents concerned by the user according to a multi-dimensional evaluation standard to generate different expectations of the job seeker; setting weights for different expectations, processing the different expectations based on the weights, and generating a final expected analysis result of the job seeker;

analyzing recruiters: setting recruitment standards, and screening big data in three ranges of the same industry, the same post type and the same standard type according to the set recruitment standards to generate a scoring system; using the scoring system to perform weight scoring on the employment requirements of the recruiting positions according to different recruitment standards; analyzing the recruiting positions based on the weight scores to generate a final expected analysis result of the recruiter; wherein the recruitment criteria comprises at least one of: the recruiter sets standards, post transverse standards and industry transverse standards by the recruiter;

wherein the user comprises the job seeker and/or the recruiter.

3. The system of claim 2, wherein the job seeker analysis model is further configured to: defining the type of the job seeker, marking the job seeker according to the type, extracting resume content of the job seeker, and acquiring a feature identifier; combining and sorting the feature identification and the big data to grade the extracted features, and matching the extracted features with corresponding recruitment posts according to different grades; the fixed level comprises a first level direct demand, a second level direct related demand, a third level direct association demand, a fourth level indirect related demand and a fifth level indirect association demand.

4. The system of claim 1, wherein the capability level analysis model is configured to at least one of:

classifying and analyzing the competence level of the job seeker: based on the working ability, skill operating ability, communication ability, execution ability, logical thinking ability, character judgment and personality judgment of the job seeker, independently splitting and confirming the ability level of the job seeker, and carrying out measurement analysis on the split and confirmed ability level of the job seeker according to a plurality of dimensions of post standard, industry standard, environment standard, skill standard and experience standard to generate an ability evaluation analysis result of the job seeker;

recruiter competency level classification and analysis: generating capability assessment analysis results for the recruiter based on recruiting posts, teams, work content, directoryworking, enterprise situation, promotion, future development.

5. The system of claim 1, wherein the dispatch output module is configured to at least one of:

automatic matching: based on at least one of: the method comprises the following steps of performing industry matching analysis, demand analysis, trend analysis, commonality analysis, general analysis and iterative analysis to generate classification and evaluation on the whole industry; and for at least one of: integrating the post content, the post advantages and the post characteristics, and matching and fitting the recruitment post and the job seeker again based on the classification and evaluation of the whole industry;

and (3) online response: generating an alternative interaction scheme through an instant interaction analysis model based on the interaction matching result and the automatic matching result, and performing alternative matching to perform online response;

6. The system of claim 1, wherein the interaction computation module is configured to at least one of:

and (3) calculating the scheduled interaction: performing operation simulation of interactive contents on a user before an interactive process occurs based on a preset scheme, and outputting interactive contents for a user simulation interactive module to simulate an interactive situation to be generated;

and (3) instant interactive calculation: and performing operation analysis on data generated in real time in the interaction process, dynamically identifying user characteristics, and performing prejudgment based on the identified user characteristics to output instant interaction content.

7. The system of claim 1, wherein the interaction computation module is configured to at least one of:

analog analysis of job seeker: generating a scheme for simulating the job seeker according to the interest analysis result, the ability level analysis result and the expected analysis result of the job seeker; performing side writing on the job seeker according to the scheme for simulating the job seeker, and determining the basic character and the psychological characteristic of the virtual role of the job seeker; performing simulation behavior analysis on the job seeker in the interaction process, and determining the behavior tendency of the job seeker;

and (3) recruiter simulation analysis: after the post standard of the recruiter is determined, information is collected and integrated according to the conditions of the recruiter enterprises, teams and leader, and a recruiter simulation scheme is generated based on the integrated data.

8. An intelligent interview method, comprising:

classifying operation is carried out on the expectation of the user through a pre-established expectation analysis model to obtain an expectation analysis result, and the ability level classification evaluation is carried out on the ability of the user through a pre-established ability level analysis model to obtain an ability analysis result;

matching the user based on the expected analysis result and the ability analysis result;

generating a pre-interaction scheme based on the relevant information of the matched users, and analyzing the interaction data generated by the matched users in the interaction process to generate an interaction response scheme;

creating a virtual role based on the pre-interaction scheme, and generating a response of the virtual role in the interaction process based on the interaction response scheme so as to perform intelligent interview;

wherein the method further comprises:

determining an approximate direction of the extension region using randomly varying probabilistic binary entropy coding and a fixed bit length of a size of a last block of the adjacent data blocks;

encoding the image based on the approximate direction of the extended region after refinement;

determining a length of a prefix of a coded syntax element in the refinement information, and refining an approximate direction of the extension region based on an approximate direction of a maximum slope among a plurality of approximate directions of an index of the prefix and an approximate direction of an angle local maximum density among the plurality of approximate directions of the index of the prefix according to the determined length of the prefix;