This project enhances customer churn prediction by combining C5 decision tree and neural network models for improved accuracy.

Accurate churn forecasting helps businesses develop effective retention strategies, crucial for maintaining growth and profitability.

• Python, NumPy, Pandas, Matplotlib
• C5.0 Decision Tree Algorithm, Artificial Neural Network (ANN)
• scikit-learn, StandardScaler, OneHotEncoder, ColumnTransformer

1. Upload the Project Files:

   • Go to Kaggle.
   • Create a new notebook.
   • Upload the necessary project files (`ai-driven-customer-churn-forcasting.ipynb, dataset CSV, etc.).

2. Dataset:

   • Ensure the dataset is uploaded to Kaggle or available via Kaggle Datasets.
   • Modify the script to load the dataset correctly from Kaggle’s environment, if necessary.

1. Data Preparation: Load and preprocess the dataset.
2. Model Building:
   • Train C5 decision tree and neural network models.
   • Combine their predictions for a final hybrid model.
3. Evaluation: Use confusion matrices and accuracy comparisons to evaluate model performance.

• Improved Accuracy: The hybrid model shows better accuracy than individual models.
• Confusion Matrix: Highlights model performance in classifying churn vs. non-churn.

The hybrid model effectively combines decision tree and neural network strengths, providing better churn predictions. Limitations include data quality, model complexity, and generalizability across different datasets.

• Project Development: niloy104

If you find this project helpful or interesting, please give it a star and follow on GitHub! It will be greatly appreciated.