Description

Machine learning program with the aim of classifying a stroke. It deals with a very delicate subject, which marks the death of 240 thousand people in Italy alone. Unfortunately, to date no one knows what caused it and how it can be avoided. To date, even having a healthy lifestyle and using drugs, we are unable to avoid the problem Collected a series of parameters of some patients, using some classification models starting from the simplest such as Perceptron and Logistic Regression (Lasso / Ridge / ElasticNet), with the use of advanced mechanisms, Skf and GridSearch, to then use regression trees like Random Forrest and ExtremGradient Boosting, models are obtained with an accuracy of up to about 95% on data never analyzed.