CS-GO-analysis
Authors : Folino Filippo Andrea, Sullazzo Teodoro (@teo-sl)
This repository contains the source code of a university project for my "Data mining" course. The task of the projet, related to CS-GO, is to predict which team will win the round. The project includes the various phases of the CRISP-DM methodology, such as:
- Business Understanding
- Data understanding
- Data preprocessing
- Modeling
- Model evaluation
Different models have been considered, some of which are:
- Decision trees
- naive bayes classifiers
- Random forest
- artificial neural networks
- SVM with SGD optimizer
- XGBoost
To each model was applied a phase of hyperparameter tuning.