Model Performance Investigator

Short Description

    Model Performance Investigator is used to analyse the performance of Machine Learning and
Deep learning models. It gives the user, basic idea how the model performance on given data.
So that user can start work on the right model.
    Current Version supports Regression and Classification type of problems with 
Visualization support.

Installation

pip install model-performance-investigator

How to use ?

* Import the package.
    from model_analyzer.ml_models import MLPredictor as ml

* Call the required function, assign it to a variable and run.

Functions ?

1. predector()  ->  Used to get Score for the Models. Returns score data
                    if error is not present else returns error message.

2. draw_plot()  ->  Used to plot Distribution plot of the Features. Return 
                    PDF file if correct dataframe send with only 'int' 
                    featurs only else returns error message.

Parameters ?

1. predector():
    * prob_type -> It is mandatory parameter.

                   TYPE : String

                   KEYS :
                    * regression
                    * classification      

                   FORMAT :
                       prob_type = ['regression']

    * data -> data must contain "train_X, train_y, test_x". train_X, train_y is for fitting
               the model and test_x is for predicting. It is mandatory.

               TYPE : List

               FORMAT :
                   data = [train_X, train_y, test_x]

    * alg_type -> Type of the algorithm. It is optional parameter. Provide the parameter
                  values in list.  
                  Default is "Linear Regression" algorithm.

                  TYPE : List

                  KEYS :
                     Regression :
                         Algorithm Name             | Parameter 
                         ---------------------------|------------------------
                         Linear Regression          | linear
                         Polynomial Regression      | polynomial
                         Redige Regression          | ridge
                         Lasso Regression           | lasso
                         ElasticNet Regression      | elasticnet
                         Random Forest Regression   | random_forest_regressor
                         XG Boost Regressor         | xgb_regressor

                    Classification :
                         Algorithm Name             | Parameter 
                         ---------------------------|------------------------
                         Logistic                   | logistic
                         Decision Tree              | decision_tree
                         Random Forest              | random_forest
                         Naive Bayes                | naive_bayes
                         SVC                        | svc
                         XGB Classifier             | xgb_classifier

                 FORMAT :
                    alg_type = ['linear', 'polynomial']


    * score_type -> Type of the score. It is optional parameter. Provide the parameter
                    values in list. Default is "r2" algorithm.

                    TYPE : List

                    1. Regression:
                       It Supports:
                        * r2
                        * explained_variance
                        * max_error
                        * neg_mean_absolute_error
                        * neg_mean_squared_error
                        * neg_mean_squared_log_error
                        * neg_median_absolute_error
                        * neg_mean_poisson_deviance
                        * neg_mean_gamma_deviance

                    2. Classification:
                        It Supports:
                        * jaccard
                        * f1
                        * neg_log_loss
                        * roc_auc
                        * accuracy
                        * balanced_accuracy
                        * average_precision

                    FORMAT :
                        score_type = ['explained_variance', 'neg_mean_poisson_deviance']


    * tune_param ->  Tuning Parameters for the model. It is an optional. By Default, 
                     this package uses some basic parameters for each models.
                     User can provide own Tunning Parameters.

                     TYPE : List of Dict

                     Template:
                         tune_param = {Name of the model: {parameters}}
                         Name of the model -> should be similar like alg_type.

                     FORMAT:
                         linear_tune_param = {
                              "fit_intercept"       : [True, False],
                              "normalize"           : [True, False],
                              "copy_X"              : [True, False],
                              "n_jobs"              : [ -1]
                            }
                        lasso_tune_param = {
                              "alpha":[1e-15, 1e-10, 1e-8, 1e-3, 1e-2, 
                                       1, 5, 10, 15, 20, 40, 50,
                                       85, 100, 300, 500, 1000
                                       ]
                            }

                         tune_param = {'linear': linear_tune_param, 
                                      'lasso': lasso_tune_param}

                     Note : User don't have to specify tunning parameters for all the
                            models used in 'alg_type'. If tunning parameters are not provided
                            then this package use default tunning parameters.

    * set_plot -> Creates Residual Plot for regression model only.
                  By default, it is Set as True.

                  TYPE : Boolean

                  FORMAT :
                     set_plot = True


EXAMPLE :
    from model_analyzer.ml_models import MLPredictor as ml

    data = [train_X, train_y, test_x]
    new_out = ml.predector('regression', data, alg_type=['linear','lasso'], 
              score_type=['r2'], tune_param='default', set_plot=True)


2. draw_plot():
    * prob_type -> Currently it supports 'regression' only. It is mandatory parameter.

                   TYPE : String

                   FORMAT :
                     prob_type = 'regression'

    * datafarme -> Used needs to send Dataframe objetc. All the features must be 
                   integers only.

    * columns -> It is optional parameter. By default all the columns are selected.
                 User can send required features alone.

                 TYPE : List

                 FORMAT :
                     columns = ['ColA', 'ColB', 'ColC']

    * plot_type -> It is optional parameter. By default 'Histograme' plot is selected.
                   User can select Histograme or Scatter plots

                   TYPE : String

                   Keys :
                       Histograme -> hist
                       Scatter    -> scatter

                   FORMAT :
                     plot_type = 'hist'


EXAMPLE :
    from model_analyzer.ml_models import MLPredictor as ml

    output = ml.draw_plot('regression', dataframe, columns =  ['ColA', 'ColB', 'ColC'],
             plot_type='hist')

Project History

    The project was started in 2019 by Srikandan Raju and Sathish Anandha.

NOTE :

    Download latest version only.