Telecom-Churn-Prediction

A classification project for forecasting the customers who are likely to churn. Random forests, XGBoost, SVM used.

Sections in the notebook:

Data Scrubbing/Cleaning: Columns Renaming & Missing Values Treatment
Exploratory Data Analysis: Correlation, Univariate & Bivariate Analysis using Visualizations
Pre Model Building Steps: Feature Enginerring and Multicollinearity Check
Model Building: Logistic Regression
Model Building: Random Forests- Deriving feature importances and building RF Model using best parameters from Grid Search CV
Model Building: SVM's- Standardizing Data and then running SVM using best parameters from Grid Search CV
Model Building: XGBoost using using best parameters from Random Search CV
Model Comparisons/Evaluation using Classification Report and ROC-AUC Curve