Installation and User Guide¶

Use the following command to install the package:

pip install dominance-analysis

The dominance-analysis pakage can then be invoked through the below code.

dominance_classification=Dominance(data=df_data,target='Target',objective=0,pseudo_r2="mcfadden",data_format=0)

The overall R-square of the complete model will be displayed as an output.

Parameters¶

The descriptions of the parameters to be passed to the Dominance class are provided below.

data - Complete Dataset, should be a Pandas DataFrame
target - Name of the target variable, it should be present in passed dataset.
top_k - No. of features to choose from all available features. By default, the package will run for top 15 features.
objective - It can take the values 0 or 1.

0 for Classification

1 for Regression

By default, the package will run for Regression.
pseudo_r2 - It can take one of the Pseudo R-Squared measures - mcfadden, nagelkerke , cox_and_snell or estrella, where default = mcfadden. It is not needed in the case of regression models (i.e. objective = 1 ).
data_format - It can take the values 0, 1 or 2.

0 when raw data is being passed,

1 when correlation matrix (correlation of predictors with target variable) is being passed,

2 when covariance matrix (covariance of predictors with target variable) is being passed.

By default, the package will run for raw data (i.e. data_format = 0). This parameter is not needed in case of Classification models.

Note: While passing a Covariance / Correlation matrix to dominance-analysis , it is advisable to pass the matrix of a dataset having 15 or lesser predictor variables.

User Guide¶

The package has the below functions that can be used for performing Dominance Analysis and coming up with visualizations that help understand the variable significance and dominance levels.

class Dominance

Function	Utility
`incremental_rsquare()`	The function will evaluate the Overall Average Incremental R-square contribution of the predictors to the R-square of the complete model.
`plot_incremental_rsquare()`	The function will plot the Incremental R-square contribution of the predictors in the form visulaizations like Bar Graph, Pie Chart and Waterfall chart.
`dominance_stats()`	The function will give the Dominance Statistics for each of the predictor variables.
`dominance_level()`	For each predictor variable, the function will clearly list out all the predictors that are dominated generally, conditionally and completely by it.
`complete_model_rsquare()`	The function will print the R-squared value of the complete model.

class Dominance_Datasets

Function	Utility
`get_breast_cancer()`	The function will fetch the UCI ML Breast Cancer Wisconsin (Diagnostic) dataset, in the form of a Pandas dataframe, to be able to use it for Dominance Analysis. The response variable in this case is continuos.
`get_boston()`	The function will fetch the Boston Housing Dataset dataset, in the form of a Pandas dataframe, to be able to use it for Dominance Analysis. The response variable in this case is binary.

You can find a more detailed information and examples regarding the package in the Official Dominance Analysis Documentation.