gdayf.core.adviserastar.AdviserAStar Class Reference

Class focused on execute A* based analysis on three modalities of working Fast: 1 level analysis over default parameters Normal: One A* analysis for all models based until max_deep with early_stopping Paranoiac: One A* algorithm per model analysis until max_deep without early stoping. More...

Inheritance diagram for gdayf.core.adviserastar.AdviserAStar:

Collaboration diagram for gdayf.core.adviserastar.AdviserAStar:

Public Member Functions
def	__init__ (self, e_c, deep_impact=3, metric='train_accuracy', dataframe_name='', hash_dataframe='')
	Constructor. More...
def	optimize_models (self, armetadata)
	Method mangaing the generation of possible optimized models. More...
Public Member Functions inherited from gdayf.core.adviserbase.Adviser
def	__init__ (self, e_c, deep_impact=5, metric='accuracy', dataframe_name='', hash_dataframe='')
	Constructor. More...
def	set_recommendations (self, dataframe_metadata, objective_column, amode=POC, atype=None)
	Main method oriented to execute smart analysis. More...
def	analysisnormal (self, dataframe_metadata, objective_column, amode)
	Method oriented to execute smart normal and fast analysis. More...
def	analysispoc (self, dataframe_metadata, objective_column, amode)
	Method oriented to execute poc analysis. More...
def	analysis_specific (self, dataframe_metadata, list_ar_metadata)
	Method oriented to execute new analysis. More...
def	analysisparanoiac (self, dataframe_metadata, objective_column, amode)
	Method oriented to execute smart normal and fast analysis. More...
def	analysisanomalies (self, dataframe_metadata, objective_column, amode)
	Method oriented to execute unsupervised anomalies models. More...
def	analysisclustering (self, dataframe_metadata, objective_column, amode)
	Method oriented to execute unsupervised clustering models. More...
def	base_specific (self, dataframe_metadata, list_ar_metadata)
	Method oriented to generate specific candidate metadata. More...
def	base_iteration (self, amode, dataframe_metadata, objective_column)
	Method oriented to select initial candidate models. More...
def	load_frameworks (self)
	Method oriented to get frameworks default values from config. More...
def	get_analysis_objective (self, dataframe_metadata, objective_column, atype=None)
	Method oriented to analyze DFmetadata and select analysis objective. More...
def	get_size_increment (self, df_metadata)
	Method oriented to analyze get increments on effort based on DF_metadata structure. More...
def	get_candidate_models (self, atype, amode, increment=1.0)
	Method oriented to analyze choose models candidate and select analysis objective. More...
def	applicability (self, model_list, nrows, ncols)
	Method oriented to select applicability of models over min_rows_limit. More...
def	priorize_models (self, model_list)
	Method managing scoring algorithm results params: results for Handlers (gdayf.handlers) More...
def	generate_vectors (self, model, normalization_set)
	Store executed model base parameters to check past executions. More...
def	is_executed (self, vector)
	Check if model has benn executed or is planned to execute. More...
def	compare_vectors (vector1, vector2)
	Compare to execution vectors. More...
def	safe_append (self, model_list, model)
	Check if model is previously executed. More...

Additional Inherited Members
Static Public Member Functions inherited from gdayf.core.adviserbase.Adviser
def	get_train_accuracy (model)
	Method get train accuracy for generic model. More...
def	get_test_accuracy (model)
	Method get test accuracy for generic model. More...
def	get_combined_accuracy (model)
	Method get averaged train and test accuracy for generic model. More...
def	get_train_rmse (model)
	Method get rmse for generic model. More...
def	get_test_rmse (model)
	Method get test rmse for generic model. More...
def	get_cdistance (model)
	Method get clustering distance for generic model. More...
def	get_train_r2 (model)
	Method get train accuracy for generic model. More...
def	get_test_r2 (model)
	Method get test accuracy for generic model. More...
Public Attributes inherited from gdayf.core.adviserbase.Adviser
	timestamp
	an_objective
	deep_impact
	analysis_recommendation_order
	analyzed_models
	excluded_models
	next_analysis_list
	metric
	dataframe_name
	hash_dataframe
	deepness
Static Public Attributes inherited from gdayf.core.adviserbase.Adviser
int	deepness = 1

Detailed Description

Class focused on execute A* based analysis on three modalities of working Fast: 1 level analysis over default parameters Normal: One A* analysis for all models based until max_deep with early_stopping Paranoiac: One A* algorithm per model analysis until max_deep without early stoping.

Definition at line 30 of file adviserastar.py.

Constructor & Destructor Documentation

__init__()

def gdayf.core.adviserastar.AdviserAStar.__init__	(		self,
			e_c,
			deep_impact = 3,
			metric = 'train_accuracy',
			dataframe_name = '',
			hash_dataframe = ''
	)

Constructor.

Parameters

self	object pointer
e_c	context pointer
deep_impact	A* max_deep
metric	metrict for priorizing models ['train_accuracy', 'test_rmse', 'train_rmse', 'test_accuracy', 'combined_accuracy'] on train
dataframe_name	dataframe_name or id
hash_dataframe	MD5 hash value

Definition at line 39 of file adviserastar.py.

Member Function Documentation

optimize_models()

def gdayf.core.adviserastar.AdviserAStar.optimize_models	(		self,
			armetadata
	)

Method mangaing the generation of possible optimized models.

Parameters

armetadata

ArMetadata Model

Returns

list of possible optimized models to execute return None if nothing to do

Definition at line 46 of file adviserastar.py.

Here is the call graph for this function:

Here is the caller graph for this function:

---

The documentation for this class was generated from the following file:

• D:/OneDrive/DayF/git/gdayf-core/src/gdayf/core/adviserastar.py