This is a university project in which we carried out several data mining tasks on a dataset containing information about tennis matches. There were several steps, initially a data understanding phase and then a feature engineering phase, with the aim of defining a new dataset containing player profiles. Then, we applied clustering analysis (k-means, DBSCAN, hierarchical clustering, EMA, X-means, C-means) to identify a potential interesting pattern among the players. Finally, we created a label to discriminate strong players from weak players and applied several classification techniques (Random Forest, AdaBoost, Neural Networks, SVM, Decision Tree, Rule Based, KNN, Naive Bayes). The project concludes with the analysis of some time series on the temperatures of some world cities. A detailed explanation showing the results and the choices made can be in the report.
With PIP
pip install -r requirements.txt
With CONDA
conda create --name <env_name> --file requirements.txt
In order to run the jupyter notebooks, first execute jupyter notebook and then run the notebooks in the order specified by the numbers in the file name.