These are the base solutions, in a variety of languages, for the 'KNN Digits Recogniser' dojo. Most include an example completed solution.
Machine Learning 101 Slides
Box Car 2D, demonstrating a genetic algorithm
The original MNIST hand drawn character dataset, which includes stats for how well various machine leaning algorithms have performed
[Kaggle] (http://www.kaggle.com/competitions) which includes an ongoing competition for digit recognisers
@brandewinder's original F# Dojo, where this originated
My Digit recognition in F# with k-nearest neighbours blog post, which steps through a sligtly more in-depth implementation than you'd typically manage in the dojo.
Python base solution and example code added May 2014 by Alex Garland @alexdgarland. These should work in anything >= Python 2.7, on Windows or Linux.
Java version made with v7. Done in IntelliJ IDEA, but you should be able to just use Main.java in whatever IDE you want.
PHP version by Alfonsas Cirtautas. Works with PHP 5.4 and up, code should be executed in cli.
Scala version by Matt Russel, is an implementation from a workshop and not a guided script like the rest.
Ruby version, only partially complete so far, by Gwen Diagram