Customer churn prediction

Description

We are trying to build a model which predicts whether a person purchases an item present in his or her cart on a e-commerce platform.

Data Exploration

1. Slice and dice data to understand each attribute
2. Generate plots to understand the distribution

Feature Engineering

1. Advanced EDA
2. Splitting category code at two levels
3. Capture User activity (session count, total activity in session, product affinity)
4. Binning hours in 4 bucket
5. Reduction in brands
6. Generate Target variable

Model selection

1. Evaluate these algorithm
   • Logistic Regression
   • Decision Tree
   • Random Forest
2. Hyperparameter Tunning (ParamGridBuilder)
3. Threshold selection

Model Inference

1. Feature Importance
2. Model implications