By Alexey Golgovskiy.
A step-by-step demonstration of a machine learning task. This is a test task that I have solved before. It is not the most optimal way to solve this task, but it was created for demonstration purposes.
Churn modelling: Predicting whether the customer will leave.
Dataset does not belong to me. (https://www.kaggle.com/datasets/shrutimechlearn/churn-modelling) Included 10,000 records (churn_modelling.csv)
95% accuracy in cross-validation for F1
- Conducted EDA
- Analyzed data
- Preprocessed features
- Trained several popular classification models for testing and comparison
- Trained the final model and fine-tuned it for the final score
- Provided conclusions and some important visualizations + feature importance