How to build a Machine Learning Platform and an example of use
Version 0.1
By Pierfrancesco Ghedini
twitter account @pierfghedini
web: http://informaticasanitaria.it
In this tutorial we will setup a platform from scratch to experiment with Python and Machine Learning tools. What you will need:
- an host enabled to run Virtual Machines
- or an host with a USB memory key of adequate dimension to store a linux distro (at least 2GB)
In order to make simple the setup we will use the "Fedora Python Classroom" an excellent linux distro with a preinstalled lot of useful tools.
The tutorial is also an introduction to simple Machine Learning techniques. In particular we will use a well known classifier called "RandomForestClassifier" to try to predict if a passenger of the famous TITANIC will survive to the shipwreck or not. As the manual says the RandomForestClassifier is "a meta estimator that fits a number of decision tree classifiers on various sub-samples of the dataset and use averaging to improve the predictive accuracy and control over-fitting."