CN110297967B - Method, device and equipment for determining interest points and computer readable storage medium - Google Patents

Abstract

The present disclosure provides a method, an apparatus, a device and a computer-readable storage medium for determining a point of interest, including: mining network resources through a pre-trained mining classifier to obtain interest points; determining an incidence relation between the interest points according to the obtained interest points; and determining the target interest points of the user according to the incidence relation among the interest points. The method, the device, the equipment and the computer readable storage medium provided by the disclosure can extract the interest points in time according to network resources and construct the association between the interest points, so that the target interest points of the user can be determined according to the known user information and the association relation between the interest points, and the content which is possibly interested by the user can be determined quickly under the condition that the iteration speed of the internet information is high.

Description

Point-of-interest determination method, apparatus, device, and computer-readable storage medium

technical field

The present disclosure relates to content push technology, and in particular, to a method, apparatus, device, and computer-readable storage medium for determining a point of interest.

Background technique

With the development of the Internet, personalized content recommendation for users is becoming more and more popular. By recommending content to users in a targeted manner, users can more quickly find interesting content and browse for users.

In the personalized recommendation process, the method adopted in the prior art is to determine the user's interest or potential interest based on the user's historical data, and then push it to the user.

However, the iteration speed of network information is relatively fast. If content is only pushed to users based on the user's historical data, it is easy to push the content related to the historical information to the user, but it is impossible to push the current newer content to the user. Therefore, this tentative recommendation method cannot accurately determine the content that the user may be interested in at present, and thus cannot accurately push the interested content to the user.

SUMMARY OF THE INVENTION

The present disclosure provides a method, apparatus, device, and computer-readable storage medium for determining a point of interest, so as to solve the problem that in the prior art, the content of the user's current interest cannot be accurately determined, and thus the content of interest cannot be accurately pushed to the user. question.

A first aspect of the present disclosure is to provide a method for determining a point of interest, including:

Mining network resources through pre-trained mining classifiers to obtain points of interest;

Determine the association relationship between the points of interest according to the acquired points of interest;

The target interest point of the user is determined according to the relationship between the interest points.

Another aspect of the present disclosure is to provide an apparatus for determining a point of interest, including:

The mining module is used to mine network resources and obtain points of interest through pre-trained mining classifiers;

an association module, configured to determine the association relationship between the interest points according to the acquired interest points;

The determining module is configured to determine the target interest point of the user according to the association relationship between the interest points.

Yet another aspect of the present disclosure is to provide a point-of-interest determination device, including:

memory;

processor; and

Computer program;

Wherein, the computer program is stored in the memory and configured to be executed by the processor to implement the method for determining a point of interest as described in the first aspect above.

Yet another aspect of the present disclosure is to provide a computer-readable storage medium on which a computer program is stored, the computer program being executed by a processor to implement the method for determining a point of interest as described in the first aspect above.

The technical effects of the method, apparatus, device and computer-readable storage medium for determining a point of interest provided by the present disclosure are:

The method, device, device and computer-readable storage medium for determining a point of interest provided by the present disclosure include: mining network resources through a pre-trained mining classifier to obtain a point of interest; and determining an association between the points of interest according to the acquired point of interest Relationship; determine the user's target POI according to the relationship between POIs. The method, apparatus, device, and computer-readable storage medium provided by the present disclosure can timely extract points of interest according to network resources, and build an association between the points of interest, so that the known user information and the association relationship between the points of interest can be used. Determine the user's target point of interest, so as to quickly determine the content that the user may be interested in when the iteration speed of Internet information is fast.

Description of drawings

FIG. 1 is a flowchart of a method for determining a point of interest according to an exemplary embodiment of the present invention;

2 is a flowchart of a method for determining a point of interest according to another exemplary embodiment of the present invention;

3 is a structural diagram of an apparatus for determining a point of interest according to an exemplary embodiment of the present invention;

4 is a structural diagram of an apparatus for determining a point of interest according to another exemplary embodiment of the present invention;

FIG. 5 is a structural diagram of a device for determining a point of interest according to an exemplary embodiment of the present invention.

Detailed ways

With the development of network technology, more and more users will obtain needed information on the network, for example, watching some hot news. In order to provide users with better services, many network platforms will actively push content to users, such as news that happened today, and hot content that users may care about.

In the prior art, the network platform obtains the user's historical data, infers the user's interest according to the historical data, selects the content that the user may be interested in from the news content or hot content, and pushes it to the user.

However, with the rapid iteration of network information, the accuracy of inferring content that may be of interest to a user based solely on historical data is low. The solution provided by the embodiment of the present invention can mine points of interest in network data with a very fast iteration speed, and determine the correlation between the points of interest, so that the target point of interest of the user can be accurately determined based on the correlation, so that based on the correlation When the target interest point pushes content to the user, it can match the content that is more in line with the user, thereby improving the push efficiency.

FIG. 1 is a flowchart of a method for determining a point of interest according to an exemplary embodiment of the present invention.

As shown in FIG. 1 , the method for determining a point of interest provided by this embodiment includes:

Step 101 , mining network resources through a pre-trained mining classifier to obtain points of interest.

The method provided in this embodiment may be executed by an electronic device with computing capability, such as a computer. The electronic device may be a background server of a network platform, and the method provided in this embodiment may be encapsulated in the server, so that the server can execute the method.

Wherein, the electronic device may be connected with the user terminal through a network, so as to push content to the user terminal. For example, a client terminal may be installed in the user terminal, the user may operate the client terminal to browse network contents, and the electronic device may push contents to the user terminal through the function of the client terminal.

Specifically, a pre-trained mining classifier may be set in the electronic device. The mining classifier can process network resources to extract points of interest that may be included. The point of interest refers to the content that the user may be interested in, such as a popular star at the moment, such as an unexpected event, etc.

Further, some points of interest already exist, such as points of interest generated months or even years ago, which can be regarded as known points of interest. There are also some points of interest that are being generated or have not yet been generated. Due to the very fast circulation of network information, many emergencies will quickly spread in the network. Therefore, new points of interest will be generated all the time. Through the pre-trained mining classifier, the network resources can be processed, and the current interest points can be obtained in time.

In practical application, the classifier can be trained according to the known points of interest and their corresponding network resources. For example, some points of interest and their corresponding network resources can be selected as samples, which can include some positive samples and negative samples. Positive samples are the correct combination of interest points and network resources, and negative samples are the wrong combination of interest points and network resources.

It is also possible to use a part of the samples as training samples and another part of the samples as test samples. Execute the classifier algorithm on the training samples to generate a classifier. Execute the classifier on the test samples to generate predictions. Determine the evaluation index according to the prediction result, and evaluate the performance of the classifier. If the classifier does not meet the standard, you can adjust the parameters in the classifier and retrain.

The network resource may be web page data, for example, a classifier capable of mining interest points in the web page data may be trained. The web resource can also be search data, for example, a classifier that can mine points of interest in a user's search data can be trained.

Specifically, the network resources in the network can be obtained, and the trained classifier can be used to calculate them, and the interest points included therein can be extracted, so that the current interest points in the network can be obtained in time.

Step 102: Determine the association relationship between the points of interest according to the acquired points of interest.

For the newly acquired point of interest, the relationship between it and other points of interest can be determined, so that other points of interest of the user can be inferred on the basis of the known points of interest of a user according to the relationship.

Among them, for many interest points in the network, there is a certain relationship between these interest points. For example, if the interest point A and the interest point B appear in pairs, it can be considered that the two have a certain relationship. For another example, some points of interest include multiple sub-points of interest, and this relationship can be determined according to an existing knowledge system. For example, the point of interest in the Qing Dynasty belongs to the history of the point of interest. The relationship can also be determined according to network data. For example, a TV program is a point of interest, and the guests participating in the TV program can be a sub-point of interest of the point of interest.

Specifically, by constructing the relationship between the interest points, an interest point map can be drawn, and each interest point has its corresponding associated interest point.

Step 103: Determine the target interest point of the user according to the association relationship between the interest points.

Further, a user's point of interest may be determined first, and other points of interest that have a relationship with the point of interest may be determined according to the association between the points of interest, and these other points of interest may be determined as the target point of interest of the user. For example, if the point of interest A and the point of interest B have an associated relationship, when it is presumed that the user is interested in A, it can be presumed that the user is also interested in B.

In practical applications, some points of interest have an inclusion relationship, and at this time, the target point of interest of the user can be determined according to this relationship. For example, if the user has a point of interest A, it may be considered that he may also be interested in the sub-points of interest of the point of interest A, and these sub-points of interest may be sent to the user, so that the user can select among them.

Among them, the user can also directly inquire about the points of interest that the user cares about, and then accurately determine a known point of interest, and then determine the target point of interest according to the known point of interest and the relationship between the points of interest.

In the method provided in this embodiment, the user's target interest point can be determined according to the relationship between the interest points, so that the content that the user cares about can be more accurately estimated based on the existing user information or data.

Specifically, the method provided in this embodiment may further include:

The push content corresponding to the user is determined according to the target point of interest, and the push content is sent to the user terminal.

Further, after the content that the user may care about is determined, the network content can be screened based on the content, for example, hot news is screened and pushed to the user.

The method provided by this embodiment can accurately determine the target interest point of the user based on the existing data, and then can push the content more matching with the target interest point to the user based on the target interest point.

The method provided in this embodiment is used to determine a user's point of interest, and the method is executed by a device provided with the method provided in this embodiment, and the device is usually implemented in hardware and/or software.

The method for determining points of interest provided by this embodiment includes: mining network resources through a pre-trained mining classifier to obtain points of interest; determining the association between the points of interest according to the acquired points of interest; and according to the association between the points of interest Identify the user's target point of interest. The method provided in this embodiment can extract points of interest in time according to network resources, and build an association between the points of interest, so that the target point of interest of the user can be determined according to the known user information and the relationship between the points of interest, and then When the iteration speed of Internet information is fast, the content that users may be interested in can be determined quickly.

FIG. 2 is a flowchart of a method for determining a point of interest according to another exemplary embodiment of the present invention.

As shown in FIG. 2 , the method for determining a point of interest provided by this embodiment includes:

Step 201: Extract webpage features in the webpage data through a first mining classifier, and determine points of interest according to the webpage characteristics.

The first mining classifier is obtained by training the known points of interest and their corresponding webpage data.

Specifically, known points of interest and their corresponding web page data can be collected in advance, and these data can be used as training data to train the classifier. The combination of correct interest points and webpage data can be used as positive samples, and the wrong combination of interest points and webpage data can be used as negative samples, and the first classifier can be accurately trained based on these two sample data.

Further, the electronic device for training the classifier may be an electronic device that executes the method provided in this embodiment, or may be other electronic devices.

The trained first classifier is stored in the electronic device that executes the method provided in this embodiment. The electronic device can scan web page data in the network, and calculate the web page data through the first classifier to obtain corresponding points of interest.

The web page data may be data content included in the web page, such as text content, picture content, link addresses, and the like.

When processing the web page data, the first classifier can extract the web page features included in the web page data, and then classify the web page features, so as to determine the point of interest.

The webpage features may include at least one of the following: page features, entry hotspots, and entry types. For each web page feature, one corresponding content can be extracted, and multiple corresponding contents can also be extracted.

Step 202 , extract the search features in the network search data through the second mining classifier, and determine the points of interest according to the search features.

Among them, the second mining classifier is obtained by training the known points of interest and their corresponding network search data.

The timing of step 201 and step 202 is not limited. At the same time, only one method in steps 201 and 202 can be used to mine points of interest, or both methods can be used to mine points of interest.

Specifically, the second mining classifier is obtained by training the known points of interest and their corresponding network search data.

Further, known points of interest and their corresponding network search data can be collected in advance, and these data can be used as training data to train the classifier. The combination of the correct interest point and the network search data can be used as a positive sample, and the combination of the wrong interest point and the network search data can be used as a negative sample, and the second classifier can be accurately trained based on these two sample data.

Network search data refers to data generated by users searching for information on the network.

In practical applications, the electronic device for training the classifier may be an electronic device that executes the method provided in this embodiment, or may be other electronic devices.

The trained second classifier is stored in the electronic device that executes the method provided by this embodiment. The electronic device can collect the search data performed by the user in the network, and calculate the search data through the second classifier to obtain the corresponding points of interest.

When processing the network search data, the second classifier may extract the search features included in the network search data, and then classify the network search features to determine the point of interest.

The search features include at least one of the following: search information, user click information, search timeliness information, and webpage content features. For each search feature, one corresponding content may be extracted, or multiple corresponding content may be extracted.

After the point of interest is mined based on steps 201 and/or 202, the association relationship between the point of interest and other points of interest may be determined. Specifically, it can be determined by any one of steps 203 , 205 , and 206 .

Step 203: Determine the co-occurrence information between the POIs in the network resource, and determine the association relationship between the POIs according to the co-occurrence information.

In one embodiment, it is considered that there may be a certain relationship between the points of interest that appear together. Therefore, the co-occurrence information among the POIs can be determined in the network resource, for example, the number of co-occurrences, the ratio of the number of co-occurrences to the number of individual occurrences, and the like.

For example, when a TV program and an actor's name often appear together, it can be considered that the two have an associated relationship. At this time, it can be considered that the TV program and the actor have a non-directional relationship.

Among them, a weight value can also be set to measure the strength of the association between the two interest points. For example, if the co-occurrence times of two points of interest exceeds a threshold number of times, it is considered that the correlation strength between the two points of interest is relatively high, and therefore, a larger weight value can be set.

Specifically, the weight value can also be obtained by calculation, for example, the number of co-occurrences between interest points is used as the weight value.

Further, after acquiring a point of interest, the relationship between it and other points of interest may be determined, and the other points of interest here may be obtained by mining based on the method of this embodiment, or may be obtained by other methods.

Step 204 , in the points of interest of the social network, determine the associated users corresponding to the users of the social network, and associate the associated points of interest of the users.

In another embodiment, the point of interest may be a social network user, such as user N. At this time, the related point of interest of this point of interest may also be determined through the social network.

There are many network users in the network, and some network users themselves may also be points of interest, such as some actors, singers, and the like. The social network users have social relationships in the social network. Therefore, the association relationship of the social network users belonging to the point of interest can be determined according to their associated users in the social network. For example, a user who is concerned with a social network user belonging to a point of interest can be regarded as an associated user of the point of interest, that is, it can also be regarded as a point of interest, and it is considered that there is an associated relationship between the two points of interest.

Wherein, when determining the associated user, the associated user may also be determined according to the interaction information between the two users. For example, if two users interact frequently and one of the users is considered a point of interest, the other user may be a related point of interest for the point of interest.

Step 205: Determine the pending superordinate concept of the point of interest according to the network resources.

Specifically, some points of interest have an upper-lower relationship. For example, a point-of-interest sports may include multiple other points of interest, such as football, basketball, and the like. Therefore, it is also possible to construct a point-of-interest association relationship with a pointing relationship based on the upper-lower relationship between the points of interest.

Further, an undetermined superordinate concept of a point of interest may be determined first.

In an implementation manner, a conceptual system may be constructed in advance according to an existing knowledge system in network resources. The concept system can include the superordinate names and subordinate nouns corresponding to each noun.

After a point of interest is determined, if the point of interest does not belong to the concept system, a pending superordinate concept of the point of interest can be determined in the concept system.

In another implementation manner, a to-be-located superordinate concept may also be determined according to the content of the webpage corresponding to the point of interest. Or according to the search data corresponding to the interest point, determine the to-be-determined superordinate concept.

Step 206 , according to the third classifier, determine whether the to-be-determined superordinate concept is the real superordinate concept of the point of interest.

In practical application, a third classifier may also be set in the electronic device to determine whether the to-be-determined superordinate concept is accurate.

Wherein, the third classifier may be obtained by pre-training. For example, the determined concept system can be used as training data, and the third classifier can also be trained according to the search distribution data of known points of interest and their corresponding superordinate concepts, and their combinations.

Specifically, the trained third classifier is set in the electronic device to determine whether the to-be-determined superordinate concept is accurate. The content of the webpage corresponding to the point of interest, the search data, etc., and the determined undetermined superordinate concept can be input into the third classifier, so that it can be confirmed.

Further, in step 205, a plurality of undetermined superordinate concepts may also be determined for a point of interest, and at this time, each pending concept may be confirmed in this step. A point of interest can have multiple real superordinate concepts, for example, "A Dream of Red Mansions" can have superordinate concepts "literature", and can also have superordinate concepts "history", "Qing Dynasty" and so on.

Step 207, if yes, then determine that the real superordinate concept and the point of interest have a pointing relationship

In practical application, if it is determined that the to-be-determined superordinate concept is the real superordinate concept of the point of interest, it is determined that the two have a pointing relationship, for example, the point of interest points to the real superordinate concept to which it belongs.

Wherein, if it is confirmed that each undetermined superordinate concept of a point of interest is not its real superordinate concept, it can be considered that the interest point does not yet have a superordinate concept. For example, in the early stage of establishing the relationship map of interest points, it may contain less information, and at this time, it may not be possible to determine the true superordinate concept in the existing data.

For points of interest without a superordinate concept, when a new interest point is added to the map, it can be identified according to the newly added content to determine whether it is the superordinate concept of these interest points.

Step 208: Determine the user's historical points of interest according to the user's historical data.

Specifically, the method provided in this embodiment may further determine the target interest point of the user according to the determined relationship between the interest points.

Further, the historical interest points of the user, that is, the content that the user was interested in once, can be determined according to the historical data of the user. This determination method can adopt the method in the prior art.

In practical applications, network resources are iterated very fast, and it is possible that the user's historical interest points do not match the current interest points. Therefore, if content is only pushed to users based on historical points of interest, it is easy to push content that is not of interest to users.

Step 209: Determine the target interest point associated with the historical interest point according to the association relationship between the interest points.

Wherein, the target interest point associated with the historical interest point may be determined based on the constructed association relationship between the interest points.

For example, if an undirected association relationship is constructed, the interest points that have a relationship with the historical interest points can be directly obtained, and these interest points can be directly used as the target interest points; according to the weight value between these interest points and the historical interest points, Some more relevant interest points are screened as target interest points, for example, a preset number of interest points with higher weight values are screened as target interest points.

For another example, if a directed association relationship is established, the upper-level interest point with the historical interest point can be obtained, which can be used as a target interest point. It is also possible to more accurately determine the interests of users. For example, the undirected association relationship can be combined, and the interest point that belongs to the upper interest point and has an undirected association relationship with the historical interest point can be used as the target interest point. For example, if the historical interest point is A ₁ , and its upper interest point is A, other interest points belonging to A can be obtained, such as A ₂ and A ₃ , assuming that A ₂ and A ₁ have an undirected relationship, and A ₃ and A ₁ do not If there is an undirected relationship, _A2 is considered to be a target point of interest.

Step 209 is a way of determining the target point of interest by the electronic device. In addition, the target point of interest may also be determined by interacting with the user.

Step 210: Send inquiry information including the first point of interest to the user terminal of the user.

Wherein, if a directional association relationship between the points of interest is established, the electronic device can interact with the user terminal based on the directional relationship to determine the target point of interest.

Specifically, a first point of interest may be determined first, for example, a historical point of interest as described above, the first point of interest may also be determined according to the user's current browsing content, or the first point of interest may be determined randomly.

Further, the electronic device may send an inquiry message to the user terminal for inquiring whether the user is interested in the first point of interest.

In practical application, the user can operate the user terminal to reply, such as replying yes or no, like or dislike and other information.

Step 211: Receive the interest result returned by the user terminal.

Wherein, after the user operates the user terminal to reply, the electronic device may receive the interest result fed back by the user terminal. The result can be either yes or no.

Step 212: Determine an actual point of interest according to the interest result, and in the association relationship, determine the target point of interest according to the pointing relationship of the actual point of interest.

Specifically, the actual point of interest can be determined according to the result of the user's reply. For example, if the user replies yes, it can be considered that the first point of interest is the user's actual point of interest. User asks.

Further, after the actual point of interest is determined, the target point of interest may be determined according to the pointing relationship of the actual point of interest in the directed association relationship. For example, if the user is interested in football, the electronic device may determine that the user is interested in sports, and then take sports as a target interest point.

In practical application, if targeted content can be pushed to users, more intensified target POIs need to be determined. Therefore, sub-POIs belonging to actual POIs can also be determined in the association relationship between directional POIs. And/or in a directed relationship between POIs, the parent POI of the actual POI is determined. For example, if the actual point of interest is sports, sub-points of interest in sports, such as basketball, football, gymnastics, swimming, etc., can be obtained. If the actual POI is basketball, the corresponding parent POI, such as sports, can also be obtained.

Among them, it is also possible to choose whether to obtain the child POI or the parent POI based on the actual POI. For example, if the actual POI only has the parent POI, the parent POI can be obtained, and if the actual POI only has the child POI , the sub-point of interest can be obtained. A parent POI here is the superordinate concept of its child POIs.

The query information including the child POI and/or the parent POI is sent to the user terminal of the user according to the child POI and/or the parent POI.

Specifically, after acquiring the POI associated with the actual POI, the user can be further inquired whether the POI is interested. Therefore, query information including the determined POI can be sent to the user terminal.

After seeing the inquiry information, the user can operate the user terminal to give feedback to the electronic device.

A second interest result returned by the user terminal is received.

After the user operates the user terminal to confirm whether he is interested in the current point of interest, the user terminal may send it to the electronic device, and then receive the second interest result.

Assuming that the second interest result is that the user is interested in the currently determined interest point, it can be used as a target interest point. The process of determining the target POI can be ended, and the steps of determining other child POIs belonging to the actual POI, and/or other parent POIs of the actual POI in the directional relationship between POIs can be continued.

Assuming that it is determined according to the second interest result that the user is not interested in the child POI and/or the parent POI, continue to determine other child POIs belonging to the actual POI in the association relationship, and/or determine the actual POI in the association relationship steps for other parent POIs.

If the user is not interested in the currently determined point of interest, the user may continue to determine other points of interest associated with the actual point of interest, and interact with the user based on the determined point of interest. By interacting with the user, the target point of interest can be determined more directly and accurately.

For example, the interaction process between the system where the electronic device is located and the user can be:

System: Hello, do you like sports news?

user: like;

System: Do you like NBA?

User: not interested;

System: Do you like football?

user: like;

System: Do you like Beckham?

user: like;

System: We will recommend to you based on your interests.

FIG. 3 is a structural diagram of an apparatus for determining a point of interest according to an exemplary embodiment of the present invention.

As shown in FIG. 3 , the device for determining a point of interest provided by this embodiment includes:

The mining module 31 is used for mining network resources through a pre-trained mining classifier to obtain points of interest;

an association module 32, configured to determine an association relationship between the points of interest according to the acquired points of interest;

The determining module 33 is configured to determine the target interest point of the user according to the association relationship between the interest points.

The device for determining a point of interest provided in this embodiment includes a mining module for mining network resources through a pre-trained mining classifier to acquire a point of interest; an association module for determining the association between the points of interest according to the acquired points of interest Relationship; a determination module, used for determining the target interest point of the user according to the association relationship between the interest points. The device provided in this embodiment can extract points of interest in time according to network resources, and build an association between the points of interest, so that the target point of interest of the user can be determined according to the known user information and the relationship between the points of interest, and then the user's target point of interest can be determined. When the iteration speed of Internet information is fast, the content that users may be interested in can be determined quickly.

The specific principle and implementation manner of the apparatus for determining a point of interest provided in this embodiment are similar to the embodiment shown in FIG. 1 , and details are not described herein again.

FIG. 4 is a structural diagram of an apparatus for determining a point of interest according to another exemplary embodiment of the present invention.

As shown in FIG. 4 , on the basis of the above embodiment, in the device for determining a point of interest provided by this embodiment, the mining module 31 includes a first mining unit 311 , which is used for:

Extracting webpage features in webpage data through a first mining classifier, and determining the interest point according to the webpage characteristics;

Wherein, the first mining classifier is obtained by training the known points of interest and their corresponding webpage data.

Optionally, the digging module 31 includes a second digging unit 312 for:

Extracting search features in the network search data by a second mining classifier, and determining the point of interest according to the search features;

Wherein, the second mining classifier is obtained by training the known points of interest and their corresponding network search data.

Optionally, the association module 32 includes a first association unit 321 for:

Co-occurrence information between the interest points is determined in the network resource, and an association relationship between the interest points is determined according to the co-occurrence information.

Optionally, the point of interest is a social network user;

The association module 32 includes a second association unit 322 for:

In the point of interest of the social network, the associated user corresponding to the social network user is determined, and the associated point of interest of the associated user is determined.

Optionally, the association module 32 includes a third association unit 323 for:

determining the pending superordinate concept of the point of interest according to network resources;

Determine whether the to-be-determined superordinate concept is the real superordinate concept of the point of interest according to the third classifier;

If so, it is determined that the real superordinate concept has a pointing relationship with the point of interest.

Optionally, the determining module 33 includes a first determining unit 331 for:

Determine the historical points of interest of the user according to the user's historical data;

A target POI associated with the historical POI is determined according to the association relationship between the POIs.

Optionally, the determining module 33 includes a second determining unit 332 for:

sending inquiry information including the first point of interest to the user terminal of the user;

receiving the interest result returned by the user terminal;

An actual point of interest is determined according to the interest result, and in the association relationship, the target point of interest is determined according to the pointing relationship of the actual point of interest.

Optionally, the second determining unit 332 is specifically configured to:

determining a child POI belonging to the actual POI in the association relationship, and/or determining a parent POI of the actual POI in the association relationship;

Send inquiry information including the child POI and/or the parent POI to the user terminal of the user according to the child POI and/or the parent POI;

receiving the second interest result returned by the user terminal;

If it is determined according to the second interest result that the user is not interested in the child POI and/or the parent POI, continue to determine other child POIs belonging to the actual POI in the association relationship, And/or in the association relationship, the step of determining other parent POIs of the actual POI.

The specific principle and implementation manner of the apparatus for determining a point of interest provided in this embodiment are similar to the embodiment shown in FIG. 2 , and details are not described herein again.

FIG. 5 is a structural diagram of a device for determining a point of interest according to an exemplary embodiment of the present invention.

As shown in FIG. 5 , the device for determining a point of interest provided by this embodiment includes:

memory 51;

processor 52; and

Computer program;

Wherein, the computer program is stored in the memory 51 and configured to be executed by the processor 52 to implement any of the above-mentioned methods for determining a point of interest.

This embodiment also provides a computer-readable storage medium on which a computer program is stored,

The computer program is executed by a processor to implement any of the point-of-interest determination methods described above.

Those of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments may be completed by program instructions related to hardware. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, the steps including the above method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk and other media that can store program codes.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, but not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: The technical solutions described in the foregoing embodiments can still be modified, or some or all of the technical features thereof can be equivalently replaced; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the technical solutions of the embodiments of the present invention. scope.

Claims (8)

1. A method for point of interest determination, comprising:

mining network resources through a pre-trained mining classifier to obtain interest points, wherein the interest points are contents in which users are interested;

determining an incidence relation between the interest points according to the obtained interest points;

sending query information including a first point of interest to a user terminal of the user;

receiving a first interest result returned by the user terminal;

determining an actual interest point according to the first interest result, and determining a first child interest point and/or a first parent interest point of the actual interest point in the incidence relation;

sending query information including the first child point of interest and/or the first parent point of interest to the user terminal;

determining whether the user is interested in the first child interest point and/or the first parent interest point according to a second interest result returned by the user terminal;

if the interest is found, the first child interest point and/or the first parent interest point are/is used as a target interest point;

if the interest is not in the interest, continuing to determine other child interest points belonging to the actual interest point in the incidence relation, and/or determining other parent interest points of the actual interest point in the incidence relation;

if the interest points do not belong to a pre-constructed concept system, determining the association relationship between the interest points according to the obtained interest points comprises the following steps:

determining a pending upper concept of the interest point in the concept system; the concept system comprises upper names and lower names corresponding to a plurality of nouns, and is constructed according to an existing knowledge system in network resources;

determining whether the undetermined upper concept is a real upper concept of the interest point according to a third classifier;

if so, determining that the real upper concept has a pointing association relation with the interest point;

if not, determining that the interest point is at the initial stage of establishing the interest point relation map, and the interest point does not have a superordinate concept; and when a new interest point is added into the relation graph, continuously determining whether the new interest point is a superior concept of the interest point.

2. The method of claim 1, wherein mining network resources to obtain points of interest through a pre-trained mining classifier comprises:

extracting webpage features in webpage data through a first mining classifier, and determining the interest points according to the webpage features;

the first mining classifier is obtained by training known interest points and corresponding webpage data.

3. The method of claim 1, wherein mining network resources to obtain points of interest through a pre-trained mining classifier comprises:

extracting search features in network search data through a second mining classifier, and determining the interest points according to the search features;

and the second mining classifier is obtained by training known interest points and network search data corresponding to the known interest points.

4. An apparatus for point of interest determination, comprising:

the mining module is used for mining network resources through a pre-trained mining classifier to obtain interest points, wherein the interest points are contents in which users are interested;

the association module is used for determining the association relationship among the interest points according to the obtained interest points;

the determining module is used for determining the target interest points of the user according to the incidence relation among the interest points;

the determining module comprises a second determining unit configured to: sending first interest point inquiry information to a user terminal of a user; determining an actual interest point according to an interest result returned by the user terminal; determining a first child interest point and/or a first parent interest point of the actual interest points in the incidence relation; sending query information including the first child point of interest and/or the first parent point of interest to the user terminal; determining whether the user is interested in the first child interest point and/or the first parent interest point according to a second interest result returned by the user terminal; if the interest is found, the first child interest point and/or the first parent interest point are/is used as a target interest point; if the interest is not in the interest, continuing to determine other child interest points belonging to the actual interest point in the incidence relation, and/or determining other parent interest points of the actual interest point in the incidence relation;

the association module comprises a third association unit, and if the interest point does not belong to a pre-constructed concept system, the third association unit is used for: determining a pending upper concept of the interest point in the concept system; the concept system comprises upper names and lower names corresponding to a plurality of nouns, and is constructed according to an existing knowledge system in network resources; determining whether the undetermined upper concept is a real upper concept of the interest point according to a third classifier; if so, determining that the real upper concept has a pointing association relation with the interest point; if not, determining that the interest point is at the initial stage of establishing the interest point relation map, and the interest point does not have a superordinate concept; and when a new interest point is added into the relation graph, continuously determining whether the new interest point is a superior concept of the interest point.

5. The apparatus of claim 4, wherein the excavation module comprises a first excavation unit to:

extracting webpage features in webpage data through a first mining classifier, and determining the interest points according to the webpage features;

the first mining classifier is obtained by training known interest points and corresponding webpage data.

6. The apparatus of claim 4, wherein the excavation module comprises a second excavation unit to:

extracting search features in network search data through a second mining classifier, and determining the interest points according to the search features;

and the second mining classifier is obtained by training known interest points and network search data corresponding to the known interest points.

7. A point of interest determination device, comprising:

a memory;

a processor; and

a computer program;

wherein the computer program is stored in the memory and configured to be executed by the processor to implement the method of any of claims 1-3.

8. A computer-readable storage medium, having stored thereon a computer program,

the computer program is executed by a processor to implement the method according to any one of claims 1-3.

Images