File size: 1,347 Bytes
3b8ac3d 257d3f1 3b8ac3d | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 | ---
title: PredictingCustomerChurn
sdk: gradio
emoji: 🚀
colorFrom: red
colorTo: yellow
short_description: A model for predicting telecom churn
---
# Predicting Telco Customer Churn using IBM dataset
This project applies machine learning techniques to predict customer churn using a dataset containing customer behavior and subscription details. The aim is to identify customers likely to leave a service and gain insights through model interpretability using SHAP values.
## 📊 Project Overview
The notebook performs the following tasks:
- **Data Preprocessing**
- Categorical encoding using LabelEncoder.
- Feature scaling using StandardScaler.
- Dropping irrelevant or low-impact features.
- **Exploratory Data Analysis (EDA)**
- Correlation analysis.
- KDE plots for feature distribution.
- Heatmap for multivariate correlation.
- **Model Building**
- **Random Forest Classifier**
- **Logistic Regression**
- **Model Evaluation**
- Classification Report
- Confusion Matrix
- Accuracy, Brier Score Loss, ROC AUC Score
- SHAP analysis for model interpretability
## 🧰 Technologies & Libraries
- Python
- Pandas
- Seaborn
- Matplotlib
- Scikit-learn
- SHAP
> **Note:** The file data.csv is the dataset got from Kaggle [telco-customer-churn](https://www.kaggle.com/datasets/blastchar/telco-customer-churn) |