Datasets
Models
Downloads: 1
90 lines (90 with data), 11.4 kB
0÷YČ3 Ń @ sz d Z d d l m Z m Z m Z m Z d d l m Z d d l m Z d d l
m Z d d l m
Z
d d l m Z d d l m Z d d l m Z d d
l m Z d d l m Z d d l m Z d d
l m Z d d l m Z d d l Z d d l Z d d l! Z! e d â Z" e d â Z# e d â Z$ d Z% d Z& d g Z' d Z( d Z) d Z* d Z+ d Z, Gd d ä d â Z- d S)z"It applies the training functions
Ú )┌Dict┌List┌TypeVar┌Any)┌_LogisticRegression)┌_LogisticRegressionCV)┌_MixedLinearModel)┌_RandomForestClassifier)┌_GradientBoostingClassifier)┌_DecisionTreeClassifier)┌_KNeighborsClassifier)┌ _NaiveBayes)┌_NeuralNetwork)┌ReadersWriters)┌ CONSTANTS)┌cross_val_scoreN┌ DataFrame┌ndarray┌ OrderedDictzMohsen Mesgarpourz-Copyright 2016, https://github.com/mesgarpour┌GPLz1.1zmohsen.mesgarpour@gmail.com┌Releasec @ sÉ e Z d Z d d e e e d d d É ä Z d d d d e e e e e d d d É ä Z e
e e e d d d
É ä Z e d d d
É ä Z e d d d É ä Z
e
e e
d d d É ä Z e e e d d d É ä Z e j d d d â e e e e d d d É ä Z d d e
e e e e d d d É ä Z e d d! d" É ä Z e e d# d$ d% É ä Z e e d# d& d' É ä Z e e d# d( d) É ä Z d S)*┌TrainingMethodN)┌ method_name┌path┌titlec C sÜ t j t j â | _ | j j t â t â | _ d | _ | | _
d | _ d | _ t
â | _ d | _ | d k rć | j | â n | j | | â d S)a│ Initialise the objects and constants.
:param method_name: the training method that will be used
(options: {'lr': Logistic Regression, 'lr_cv': Logistic Regression with Cross-Validation,
'mlm': Mixed Linear Model, 'rfc': Random Forest Classifier, 'gbc': Gradient Boosting Classifier,
'dtc' Decision Tree Classifier, 'knc': K-Nearest Neighbors Classifier, 'nb': Multinomial Naive Bayes,
'nn': Multi-Layer Perceptron (MLP) Neural Network}).
:param path: the directory path of the saved trained model file, using this application (if applicable).
:param title: the file name of the saved trained model file, using this application
N)┌logging┌ getLoggerr ┌app_name┌_TrainingMethod__logger┌debug┌__name__r ┌ _TrainingMethod__readers_writers┌_TrainingMethod__methodr ┌ model_labels┌ model_train┌dict┌
model_predict┌model_cross_validate┌ _TrainingMethod__init__method┌load)┌selfr r r ę r+ ˙LC:\Users\eagle\Documents\GitHub\Analytics_UoW\TCARER\Stats\TrainingMethod.py┌__init__5 s
zTrainingMethod.__init__)r r# r$ r& r' c C sh | j j d â | d k r+ t â | _ n | d k rF t â | _ nň | d k ra t â | _ n╩ | d k r| t â | _ n» | d k rŚ t â | _ nö | d k r▓ t â | _ ny | d k r═ t â | _ n^ | d k rŔ t
â | _ nC | d
k rt â | _ n( | j j t
d t | â â t j â | | _ | | _ | d k rRt â n | | _ | | _ d S)
a! Initialise the selected training method.
:param method_name: the training method that will be used
(options: {'lr': Logistic Regression, 'lr_cv': Logistic Regression with Cross-Validation,
'mlm': Mixed Linear Model, 'rfc': Random Forest Classifier, 'gbc': Gradient Boosting Classifier,
'dtc' Decision Tree Classifier, 'knc': K-Nearest Neighbors Classifier, 'nb': Multinomial Naive Bayes,
'nn': Multi-Layer Perceptron (MLP) Neural Network}).
:param model_labels: the features names to be inputted into the model.
Note: the order of features will be preserved internally.
:param model_train: the training model.
:param model_predict: the prediction outputs.
:param model_cross_validate: the cross-validation model.
z Initialise the training method.┌lrZlr_cvZmlm┌rfcZgbc┌dtcZknc┌nb┌nnz - Invalid training method: N)r r r r" r r r r
r r r
r ┌errorr ┌str┌sys┌exitr# r$ r% r& r' )r* r r# r$ r& r' r+ r+ r, Z__init__methodQ s2
z TrainingMethod.__init__method)┌features_indep_df┌feature_target┌kwargs┌returnc K sT | j j d â t | j j â | _ | j j | | j | | j | Ź | _ | j S)a@ Perform the training, using the selected method.
:param features_indep_df: the independent features, which are inputted into the model.
:param feature_target: the target feature, which is being estimated.
:param kwargs: the training method's argument.
:return: the trained model.
zTrain.) r r ┌list┌columns┌valuesr# r" ┌trainr$ )r* r7 r8 r9 r+ r+ r, r> s
zTrainingMethod.train)r: c C s2 | j j d â | j j | j | j d d g â S)zAPlot the tree diagram.
:return: the model graph.
zPlot.┌True┌False)r r r" ┌plotr$ r# )r* r+ r+ r, rA Ć s zTrainingMethod.plotc C s# | j j d â | j j | j â S)zN Produce the training summary.
:return: the training summary.
zSummarise training model.)r r r" ┌train_summariesr$ )r* r+ r+ r, rB ľ s z TrainingMethod.train_summaries)r7 ┌ sample_namer: c C sA | j j d â | j j | j | | j â | j | <| j | S)a= Predict probability of labels, using the training model.
:param features_indep_df: the independent features, which are inputted into the model.
:param sample_name: the sample to predict(e.g. 'train', 'test', 'validate').
:return: the predicted probabilities, and the predicted labels.
zPredict.)r r r" ┌predictr$ r# r& )r* r7 rC r+ r+ r, rD Ł s &zTrainingMethod.predict)r8 rC r: c C s; | j j d â | | j | d <| j j | j | | â S)a roduce summary statistics for the prediction performance.
:param feature_target: the target feature, which is being estimated.
:param sample_name: the sample to predict(e.g. 'train', 'test', 'validate').
:return: the prediction summaries.
zSummarise predictions.┌target)r r r& r" ┌predict_summaries)r* r8 rC r+ r+ r, rF ę s z TrainingMethod.predict_summariesr g═╠╠╠╠╠?gÜÖÖÖÖÖę?)r8 rC ┌cutoffsr: c C sB | j j d â | | j | d <| j j | j | d | | â S)ai Produce a summary statistics table for a range of cut-off points.
:param feature_target: the target feature, which is being estimated.
:param sample_name: the sample to predict(e.g. 'train', 'test', 'validate').
:param cutoffs: a list of risk cut-off points.
:return: the summary statistics table for the cut-off points.
zSummarise predictions.rE ┌score)r r r& r" Z predict_summaries_cutoffs_table)r* r8 rC rG r+ r+ r, ┌ predict_summaries_risk_bandsÁ s
z+TrainingMethod.predict_summaries_risk_bands┌neg_mean_squared_errorÚ
)r7 r8 ┌scoring┌cvr: c C sB | j j d â t | j | | j | d | d | â| _ | j S)aČ Evaluate the model by performing cross-validation.
:param features_indep_df: the independent features, which are inputted into the model.
:param feature_target: the target feature, which is being estimated.
:param scoring: the scoring method (default: 'neg_mean_squared_error').
:param cv: the cross-validation splitting strategy (optional).
:return: the cross-validation summary
zCross-ValidaterL rM )r ┌infor r$ r# r' )r* r7 r8 rL rM r+ r+ r, ┌cross_validate─ s (z TrainingMethod.cross_validatec C s | j S)zhProduce a summary of the applied cross-validation
:return: the cross-validation summary
)r' )r* r+ r+ r, ┌cross_validate_summariesÍ s z'TrainingMethod.cross_validate_summaries)r r c C sw | j j d â t â } | j | d <| j | d <| j | d <| j | d <| j | d <| j j | | d | âd S) aZ Save (pickle) the training model, as well as predictions and cross-validations.
Note: summaries statistics won't not saved.
:param path: the directory path of the saved trained model file, using this application (if applicable).
:param title: the file name of the saved trained model file, using this application.
z Saving modelr r# r$ r& r' ┌objectsN)
r rN r% r r# r$ r& r' r! ┌save_serialised)r* r r rQ r+ r+ r, ┌
save_model▄ s
zTrainingMethod.save_modelc C sw | j j d â t â } | j | d <| j | d <| j | d <| j | d <| j | d <| j j | | d | âd S) ad Save (pickle) & compressthe training model, as well as predictions and cross-validations.
Note: summaries statistics won't not saved.
:param path: the directory path of the saved trained model file, using this application (if applicable).
:param title: the file name of the saved trained model file, using this application.
z Save model.r r# r$ r& r' rQ N)
r r r% r r# r$ r& r' r! ┌save_serialised_compressed)r* r r rQ r+ r+ r, ┌save_model_compressedÝ s
z$TrainingMethod.save_model_compressedc C sá | j j d â | j j | | â } y@ | j d | d d | d d | d d | d d | d â Wn4 f k
rŤ | j j t d | â t j â Yn Xd S) a Load (unpickle) the model, which was saved using this application.
:param path: the directory path of the saved trained model file, using this application (if applicable).
:param title: the file name of the saved trained model file, using this application
z Load model.r r# r$ r& r' z' - Invalid field(s) in the model file: N) r r r! ┌load_serialisedr( r3 r r5 r6 )r* r r rQ r+ r+ r, r) ■ s
zTrainingMethod.load)r ┌
__module__┌ __qualname__r4 r- r r r ┌ NumpyNDArrayr( ┌PandasDataFramer> rA rB rD ┌CollectionsOrderedDictrF ┌np┌arangerI ┌intrO rP rS rU r) r+ r+ r+ r, r 4 s2 $*
0$r ).┌__doc__┌typingr r r r ZStats._LogisticRegressionr ZStats._LogisticRegressionCVr ZStats._MixedLinearModelr Z Stats._RandomForestClassifierr Z!Stats._GradientBoostingClassifierr
Z Stats._DecisionTreeClassifierr ZStats._KNeighborsClassifierr ZStats._NaiveBayesr
ZStats._NeuralNetworkr ┌ ReadersWriters.ReadersWritersr ┌Configs.CONSTANTSr Zsklearn.model_selectionr ┌numpyr\ r5 r rZ rY r[ ┌
__author__┌
__copyright__┌ __credits__┌ __license__┌ __version__┌__maintainer__┌ __email__┌
__status__r r+ r+ r+ r, ┌<module> s8 "