Project: Bank Marketing

Objective

Find out those customers who are willing to subscribe the deposit.

Challenge(s) and Solution(s)

Several challenges include bad performance of simple logistic Regression and clustering interpretation.

Solutions include add regularization such as lasso and ridge and grid_search after fix the imbalance of the target variable (too many 0 but a few 1). For clustering the segmentation is not ideally obvious but we are still able to get some insights.

Tool(s)

Python

↖ Click to view on GitHub

Required Packages

1.Logistic Regression Packages

2.Clustering Packages

Classification-Logistic Regression Model Result

Customers Segmentation-Clustering

1.Model Result of Clustering

2.Conversion Rate for Each Clustering

Note: you may check Github for additional methods including Random Forest and XGBoost