Multimedia Knowledge and Social Media Analytics Laboratory

Efficient dimensionality reduction and classification software released

×

Error message

Deprecated function: The each() function is deprecated. This message will be suppressed on further calls in menu_set_active_trail() (line 2405 of /var/www/mklab/public_html/includes/menu.inc).

We just released software (executables and sources) for our accelerated Kernel Subclass Discriminant Analysis (AKSDA) method and its combination with SVM classifiers. This is an efficient dimensionality reduction and classification method, for very high-dimensional data.

AKSDA is a new GPU-accelerated, state-of-the-art C++ library (also provided as command-line executable) for supervised dimensionality reduction and classification, using multiple kernels. It greatly reduces the dimensionality of the input data, while at the same time it increases their linear separability. Used in conjunction with linear SVMs, it achieves state-of-the-art classification results, consistently higher than Kernel SVM approaches, at orders-of-magnitude smaller training times.