Named Entity Recognition

In this project a BI-LSTM model are applied to a GeorgeTown University Multilayer Corpus (GUM). Furthermore, there is also a Topic Modelling analysis made via LDA

Blood cells cancer classification

The aim of this project is to classify 4 types of Cells, where one of them is benign and the other three are Malignant . The deep learning architectures adopted are CNN, in particular the AlexNet and VGG-16.

A Bayesian Approach to contrast CO2 Emission

The aim of this project is to analyze how the count of CO2 emission are affected by some variable (i.e Cylinders, Engine size, Transmission). The models adopted are the Bayesian Model, in particular the Linear Regression,Poisson, Hierarchical model.

Bitcoin Network

The aim of this project is to spot communities between Bitcoins transactions. The algorithm adopted is Walk-Trap, that works well with directed graph

GLM for analyzing the Trend of Bike sharing

The aim of this project is to analyze how the count of rented bikes are affected by some variables (i.e Temperature, Weather conditions). The models adopted are GLMs, in particular the ones that deal with count data.