CISUC

Fast-Decision SVM Ensemble Text Classifier Using Cluster Computing

Authors

Abstract

Text classification is a complex ubiquitous task that involves processing huge amounts of data. Nowadays people focus their attention on systems that can promptly and correctly provide tools to automatically categorize texts, like web pages or emails.
Our approach presents a fast-decision Support Vector Machine (SVM) text classifier distributing the task in a cluster environment and conjugating a set of classifiers with an ensemble strategy. Results with Reuters-21578 data set show the potential improvement in processing time and classification performance.

Subject

Text mining; SVM;GRID Computing

Related Project

CATCH - Inductive Inference for Large Scale Data Bases Text CATegorization

Conference

ICNPSC\'06, August 2006


Cited by

No citations found