Telecom Customer Churn Prediction Model (ANN & XGBoost)
This project was part of an open-ended data analysis project from my DS 740 Data Mining & Machine Learning class (UWLAX MS in Data Science)
I chose to analyze an Telecom Customer Churn data set (from Kaggle) and build XGBoost and ANN models that would classify which customer would churn.
Data Preperation & Cleaning (step 1):
The initial part of .rmd file involves exploratory data analysis, data cleaning, and preperation to ensure the quality of analysis and prepare for fitting to our two models.
Modeling (step 2):
Step 2 involves fitting the data to a ANN model and XGBoost Forest Model and fine tuning for the parameters required in both models. This step includes cross-validation and an outer fold validation conducted on both models. Here we find the best model to classify which customers churn.
Analysis (step 3):
Step 3 involves identifying the important variables and examining their relationships to the response variable.