The Random Forest is ensemble method from SkLearn library, it shows how appropriated columns values to feature importance ratios. So, each individual test has variable of conditions is based on a ratio importance that help determine the predicted result.

It is the based logic that outlines the mathematical process in the final decision and/or prediction. The CliffsNotes of Machine Learning’s little black box.

This Random Forest read in 31 Columns from the Credit DataOriginal.csv to classify. The RandomForestClassifier from SkLearn fits the default.data (less the DEFAULT column) with default.target ( just the DEFAULT column).

In order to analyze the list, a Data Frame was created displaying Feature Importance and the corresponding rating.

Then, the iloc of the top ten was utilized while sorting by descending order.

This returned the Top Ten Importance Features that were used to revise specific columns in a new data frame. The 10 columns along with the DEFAULT column became the Credit_Data_Revised.cvs used in the Neural Network Model.

Automatically, the RandomForestClassifier calculates the feature importance rating.

The rating is the weighted value of each Feature (column) in the decision process.

By sorting the feature importance in descending order, the highest feature rating lists 1–30.