VarExpPower

VarExpPower

The function describes the data in terms of explanatory power.

Parameters:

• y_train: pd.Series()
  The series of binary dependent variable of train set.
• x_train: pd.DataFrame()
  The dataframe of all independent variables of train set.
• y_test: pd.Series()
  The series of binary dependent variable of test set.
• x_test: pd.DataFrame()
  The dataframe of all independent variables of test set.
• discriminatory: str {'ttest', 'kruskal'}, default = 'ttest'
  Which test to perform to conduct the test to compare samples means. The default is 'ttest'.
• vif: bool, default = True
  Calculate the Variance inflation factor or not. If False VIFs will not be calculated.
• individual_accuracy: str {'gini', 'auc', 'f1_score'}, default = 'gini'
  Calculates individual accuracy in pairwise regression.
• check_sample: str {'test', 'train'}, default = 'test'
  Data sample to perform individual accuracy test.

Outputs:

• final_data: pd.DataFrame()
  Pandas DataFrame with all variables analysis.

Exceptions:

• TypeError:
  Raised if x_train parameter is not a pandas DataFrame object.
  Raised if x_test parameter is not a pandas DataFrame object.
  Raised if y_train parameter is not a pandas Series.
  Raised if y_test parameter is not a pandas Series.
  Raised if vif parameter is not logical.
  

• ValueError:
  Raised if discriminatory parameter is not in ('ttest', 'kruskal').
  Raised if individual_accuracy parameter is not in ('gini', 'auc', 'f1_score').
  Raised if check_sample parameter is not in ['train', 'test'].
  

Example:

import pandas as pd
from combat.short_list import VarExpPower

# Sample input data
y_train = pd.Series([0, 1, 0, 1, 0])
x_train = pd.DataFrame({'A': [1, 2, 3, 4, 5], 'B': [6, 7, 8, 9, 10]})
y_test = pd.Series([0, 1, 0, 1, 0])
x_test = pd.DataFrame({'A': [1, 2, 3, 4, 5], 'B': [6, 7, 8, 9, 10]})
discriminatory = 'ttest'
vif = True
individual_accuracy = 'gini'
check_sample = 'test'

# Calculate variable explanatory power
result = VarExpPower(y_train, x_train, y_test, x_test, discriminatory, vif, individual_accuracy, check_sample)
print(result)