Machines learning training exercises

Machine learning algorithms

Introduction

This repo contains exercises covering the main practical implementation of machine learning

Supervised learning:

non-classification algorithms

• linear regression
• linear regression with polynomial features

classification algorithms

• logistic regression
• neural networks
• support vector machine (with gaussian kernel)

Unsupervised learning: clusterisation

• K-means clustering algorithm (K = choice of clusters)
• anomaly detection algorithm: Gaussian and multiVariateGaussian (with threshold epsilon)

Dimensionality reduction algorithms: (accelerate future processing)

• Principal component analysis

Reccomandation algorithm:

• Collaborative filtering

Learning curves: Bias vs Variance

Languages/Platforms/Tools

• Matlab or GNU octave

Learning Outcomes

This was my first challenge in Matlab / GNU Octave. The goal was to hae an understanding of pratical applications of some machine learning algorithm. I think I went beyond basic understanding as I did all the necessary mathematical demonstration (available on stats.stackexchange.com).

Instructions

Install Ocatve or Matlab if you don't have it already (using brew on OSX or directly from the websites: http://www.gnu.org/software/octave/download.html)