Telecom Churn Prediction with Logistic Regression

This repository analyzes telecommunication customer churn using a Logistic Regression model built on data from Kaggle. The project focuses on:

• Data Exploration and Cleaning: Cleaning, visualizing, and understanding the churn dataset.
• Feature Engineering: Creating and transforming features for better model performance.
• Multicollinearity Analysis: Identifying and addressing highly correlated features to improve model robustness.
• Influential Points Detection: Examining data points potentially impacting model predictions.
• Logistic Regression Model Development: Training and evaluating a Logistic Regression model for predicting customer churn.

Key Points:

• Data Source: Kaggle's telecommunication churn dataset.
• Model Accuracy: 81% accuracy on the test set.
• Methodology:
  • Exploratory Data Analysis (EDA)
  • Feature Engineering
  • Multicollinearity Analysis (e.g., gvif)
  • Influential Points Detection
  • Logistic Regression Model Training and Evaluation

Contribution:

Feel free to fork and contribute to this project. Any insights or improvements are welcome!

Disclaimer: This is a basic implementation for educational purposes only. The code and approach may require further refinement for real-world deployment.