davidbroadwater / udacity-mlnd-capstone-submission

Gradient Boosted Trees for the Prediction of Professional Tennis Matches

Capstone project for Udacity Machine Learning Nanodegree

Packages/Software

All code was written in a Jupyter notebook (MLND Capstone.ipynb) running Python 3.7. A list of the libraries and packages is provided below, along with links for the less well known packages. All packages listed can be installed via pip. Note that n_jobs has been set to -1 (max number available) in most cases. On my 2012 4-core iMac, the entire notebook takes a couple hours to run (primarily due to two ~40 minute parameter tuning steps; YMMV).

Packages:

• pandas
• numpy
• matplotlib
• sklearn
• time
• pandas_profiling
• xgboost
• skopt

Data

All of the data required for this project is located in ATP.csv.

Data Profiles

Data profiles are generated at various stages of the data processing pipeline as part of the notebook, and can be opened in a browser.