Machine Learning experiments

This repository contains Jupyter Notebooks with explanations and implementation of some Machine Learning, Statistics and Data Analysis methods. Each argument is firstly described and theoretically explained and then implemented without using any external library.

Concepts are more important than efficiency.

Elements of Statistical Learning

I provide my implementation of the methods described in the famous book The Elements of Statistical Learning. I try to follow the chapter progression as much as possible.

Please remember that this code and notebooks are purely for learning purposes and therefore do not constitute a valuable source of information and could contain errors.

• (2) Overview of Supervised Learning
  • (2.3) Least Squares and Nearest Neighbors
  • (2.4) Statistical Decision Theory

Classical Machine Learning

Supervised Learning

• Linear Regression: both frequentist and bayesian approach
• Linear Classification: least squares
• Naive Bayes

Markovian Models

• Markov Chains
• Hidden Markov Model

Dimensionality reduction

• Principal Component Analysis

Statistics

• Polynomial Curve fitting: evaluation of models of several degree
• Monte Carlo Method

Neural Networks

• Restricted Boltzmann Machine

Tensorflow tutorial

Some experiments I have done while learning the Tensorflow Machine Learning library.

The code library

In the src folder there are some utility functions that I use in the various notebooks, like data generation utilities, pre-processing and so on.