Alessandro Sottile's Projects
This section contains some useful scripts for Bayesian inference obtained by the course of bayesian statistic of the data science master degree in Sapienza University
During the study, experiments were performed on a dataset of tweets regarding the price of Bitcoin. Four separate methods were used to predict the price of Bitcoin based on the average sentiment of tweets on a given day.
The aim of this work is to build a model capable of classifying diseases in corn leaves. The classes are four: Common Rust, Gray Leaf Spot, Blight, and Healthy. Three different CNN-based models are employed, with the introduction of early exit layers in the last one.
The purpose of this project is to create a football data management system. it is created using both SQL and NOSQL
The objective of this study is to explore Machine and Deep Learning techniques for the classification of orthodontic images according to the correct dental alignment treatment.
The project examined the European football market network of the top 7 leagues over the years. The process involved data extraction from Transfer Markt by web scraping, followed by graph analysis and temporal comparison.
The first part of the task involves implementing the fundamental building blocks of a transformer architecture. In the other section, however, we applied that architecture to predict human trajectories. In this part also we "played" with the various hyperparameters to analyze changes in the output of the model.
An image description model was introduced in the project with an architecture encoder/decoder. A ViT pretrained on ImageNet21k was used for the encoder and a RoBERTa model pretrained on English texts for the decoder, both of which were further trained on the Flickr8k dataset.
The goal is to answer some research questions about instagram data that may help to discover and interpret meaningful patterns in data and eventually understand how a user behaves on this social network.
Il lavoro analizza la percezione dei cittadini del Sud-Est Asiatico verso Cina e USA, considerando l'effetto dell'accesso a internet sui loro punti di vista. Utilizzando i dati dell'Asian Barometer Survey, si esplora il contesto socio-politico e storico dell'area, focalizzandosi sulle dinamiche economiche e geopolitiche attuali.
The study focuses on automating the assessment of access to bank credit using Bayesian and Frequentist Logistic Regression models, leveraging data from the "Loan Eligible" dataset on Kaggle. Metrics demonstrate good accuracy, with a particular emphasis on recall to mitigate default risk.
This repository refers to my master's thesis in the Data Science at Sapienza University. The objective is to build a model in order to perform the task of 3D fragment matching. Given two 3D scans of fragments, the model must predict whether they are adjacent or not.
Config files for my GitHub profile.