Nuclass is a tool that was created to help you train and validate classification type networks including the multilayer perceptron (MLP), functional link network, piecewise linear network, and nearest neighbor classifier.
The self organizing map (SOM) and K-Means clustering are also included. Fast pruning algorithms create a nested sequence of different size networks, to facilitate structural risk minimization.
User-supplied txt-format training data files, containing rows of numbers, can be of any size. Example training data is also provided. Fast VB Graphics for network classification error and SOM cluster formation are included.
Nuclass is highly automated and requires very few parameter choices. Advanced features include network sizing and feature selection. Training data can be compressed using the discrete Karhunen-Loeve' transform (KLT).
What's New in This Release: [ read full changelog ]
· Faster algorithms
· Improved feature selection