The purpose of this project is to develop the whole pipeline that is required to develop machine learning models that can predict churn of wireless telephony customers of a telecom in Latin America.
Together with this report we provide two Jupyter notebooks:
- Data PreProcessing and Exploration.ipynb : This notebook contains the code for the data prep-processing, exploration, and visualization.
- Model Testing and HyperParameter Tunning.ipynb : This notebook shows the training phase of our pipeline. It also includes the code for hyper-parameter tuning of several models and our feature selection strategies.
- An anonymized train and test dataset.
- A report with insights about the ML project
- Python 3.5
- Scikit-Learn
Just Launch the Jupyter Notebooks
- Roger Granda
- Maria Andrea Martens
- Alexandre Pitsaer
- This project was developed for the Advance Analytics In Business Course in KU Leuven
This repository has a report in pdf (report.pdf) with the process and insights about the project development. It was handed together with the code.