Machine Learning Small Projects

Three small, self-contained projects in optimisation, time-series forecasting and supervised learning, written during my studies at Bocconi University. Each lives in its own folder with its own README.

Project	What it does	Techniques	Entry point
ksat-simulated-annealing	Solves random K-SAT instances and locates the algorithmic threshold of 3-SAT	Simulated annealing, Metropolis MCMC	run_ksat.py
life-expectancy-arima	Forecasts a life-expectancy time series under several model orders	ARIMA (statsmodels)	arima_forecast.py
player-value-xgboost	Predicts football players' market value from their attributes	XGBoost, per-position models, 5-fold CV	player_value_xgboost.ipynb

Setup

Python 3.10 or newer.

git clone https://github.com/tommypalam/machine-learning-small-projects.git
cd machine-learning-small-projects
pip install -r requirements.txt

Data availability

Only the K-SAT project runs out of the box: it generates its own random instances. The ARIMA and XGBoost projects read CSV files that are not included in this repository, and their scripts still point at the original absolute Windows paths. Each project's README says which files are needed and which lines to change.

Repository layout

machine-learning-small-projects/
├── ksat-simulated-annealing/
│   ├── KSAT.py                  # K-SAT problem class (cost, moves, delta cost)
│   ├── SimAnn.py                # generic simulated-annealing solver
│   ├── run_ksat.py              # solve one instance
│   └── threshold_analysis.py    # solving probability and threshold experiments
├── life-expectancy-arima/
│   └── arima_forecast.py        # ARIMA fits and 10-step forecasts
├── player-value-xgboost/
│   └── player_value_xgboost.ipynb
├── requirements.txt
└── README.md