Tennis · Research
Machine Learning Based Playstyle Classification for NCAA Tennis Players
Active2025-26Updated Jan 3, 2026
Project Overview
A transfer-learning tennis playstyle classifier that turns point-by-point match charting into player archetype scores, using professional data to label styles and UCLA match data to generate radar-plot profiles.
Problem & Objective
Coaches and scouts lack a systematic way to categorize player playstyles from match statistics, relying instead on subjective observation.
Analytical Approach
The team trains classification models on match-level features to assign players to playstyle archetypes such as baseliner, serve-and-volleyer, and counter-puncher.
Key Result & Impact
The framework produces interpretable playstyle labels validated against expert ratings and surfaces dominant archetypes across the NCAA landscape.
Tech Stack & Tools
PythonGMMRandom Forest
Full paperOpen in new tab