Alex Lucchini's repositories
Stochastic-Multi-Armed-Bandits
This project consists in an empirical comparative analysis of the following stochastic multi-armed bandits algorithms: Epsilon-Greedy, SoftMax, UCB-1, UCB-V, UCB-KL, UCB-MOSS, Bayes-UCB, and Thompson Sampling. Their performance is evaluated under different metrics after multiple simulations. A peer-to-peer situation is also analysed.
OPGEE_VBA_ML
In this project it is possible to find code for the automation of VBA code through Python, and for the data analysis and prediction through ML algorithms of C02 emissions.
Leadership-Connect-NLP-Computer-Vision-Web-Scraping
In this project it is possible to find code for the text classification and topic modelling of judiciary texts, for the automated download of images from both Google Images and LinkedIn, for the panelization of large datasets, and for the comparison between people's faces.
The-role-of-context-in-classification-tasks-with-CNN
GitHub repo investigating how mimicking human contextual understanding can enhance CNNs in object classification. Uses iNaturalist dataset to study context's impact on CNN performance.
Finance-Big-Data
In this folder it is possible to find code and results for the analysis of the Capital Asset Pricing Model (CAPM) and Markowitz Portfolio Theory.
Computational-Statistics
In this project it is possible to find the code and results for acceptance rejection sampling, importance sampling, Monte Carlo and Markov Chain Monte Carlo (MCMC) simulations.