This software is obsolete. Use PyMML.
Little Snob is a library for unsupervised classification of a set of data into classes (i.e. data mining). It can handle continuous or discrete data, and is capable of determining the optimal number of classes to use on its own.
This software is in an early stage of development at the moment. It provides most of the features of "Vanilla Snob", but has not yet been extensively tested. Checking results against the original Snob is suggested. The current version is quite usable though. Surprisingly (for a Python program), it is also fairly fast, as all the heavy lifting is done by the Numerical Python library.
Little Snob has a number of advantages over the C and Fortran versions of Snob:
- Simpler, more readable code. Easier to extend or experiment with.
- Easy integration into Python software, just "import snob". For example, if your data is stored in an unusual format, it is easy to write a simple python script to read it in and then invoke Little Snob.
The Monash Data Mining Centre has been experimenting with and extending Little Snob, their version is available via anonymous CVS: