Rainbow is a C program that performs document classification usingone of several different methods, including naive BAYES, TFIDF/Rocchio,K-nearest neighbor, Maximum Entropy, Support Vector Machines, Fuhr sProbabilitistic Indexing, and a simple-minded form a shrinkage withnaive BAYES.
The Molgedey and Schuster decorrelation algorithm, having square mixing matrix and no noise . Truncation is used for the time shifted matrix, and it is forced to be symmetric . The delay Tau is estimated .
The number of independent components are calculated using BAYES Information Criterion (BIC), with PCA for dimension reduction.
BAYESerr - Computes the BAYESian risk for optimal classifier.
% BAYEScln - Classifier based on BAYES decision rule for Gaussians.
% BAYESnd - Discrim. function, dichotomy, max aposteriori probability.
% bhattach - Bhattacharya s upper limit of mean class. error.
% pBAYEScln - Plots discriminat function of BAYES classifier.
ApMl provides users with the ability to crawl the web and download pages to their computer in a directory structure suitable for a Machine Learning system to both train itself and classify new documents. Classification Algorithms include Naive BAYES, KNN