Here is a summary and breakdown of this project:
Summary:
The project is a machine learning-based approach to predicting telecom customer churn. The goal is to develop a model that can accurately predict which customers are at risk of churning, allowing the company to take proactive measures to retain them.
Modules:
-
Data Preprocessing: This module involves cleaning, transforming, and aggregating the data into a format suitable for modeling.
-
Feature Engineering: This module involves extracting relevant features from the data, including label encoding and one-hot encoding categorical variables.
-
Model Selection and Evaluation: This module involves selecting and evaluating appropriate machine learning algorithms for the task, including logistic regression, decision trees, random forests, gradient boosting, and K-nearest neighbors.
Tech Stack:
• Programming Language: Python
• Libraries: Pandas, NumPy, Matplotlib, Seaborn, Scikit-learn, CatBoost, XGBoost, LightGBM
• Tools: Jupyter Notebook