This repository hosts the final project (Pokémon_project_G13) for Statistical Learning exam, held by Professor Pierpaolo Brutti, as part of the Master’s degree in Data Science at Sapienza University of Rome.
The exam in Statistical Learning consisted of three homeworks and a final project. Instead of the final assignment, students had the option to participate in a hackathon.
This repository details the final project component of the course.
Directly from the course moodle page here you can see the project guidelines:
Our group opted for a Data-over-Method type of project, where the primary objective was to employ statistical learning techniques to determine the key characteristics that render a Pokémon competitively viable.
We constructed a dataset from the latest competitive Pokémon format (VGC24) and applied various predictive models:
Support Vector Machine.RRandom Forest.RK-Prototypes.RNeural Network.ipynb
This approach is designed to provide valuable insights that assist both novice and professional players in building effective competitive teams. Additionally, the project involved the creation of competitive Pokémon teams for each model tested (results.csv), using an algorithm named Pesca Pokémon, detailed in Chapter 4 of the project documentation.
These teams were subsequently tested on the Pokémon Showdown website to practically evaluate each team's performance in real-world competitive scenarios.
This project (plus the homeworks) received a perfect score of 30 out of 30 on the final exam. Feel free to use it as a reference if you are planning to take the exam in the upcoming years.
Please do not hesitate to contact me if you need further explanations or encounter any issues with the materials.
