Chinese
English
Associate Research Fellow  |  Chang, Fu  
 
contact
vita
education
experience
interests
descriptions
invited_talk
invited_visit
publications
patents
lab (New window)
 
 
 
 
 
Research Descriptions
 

        I have been focusing my research on various pattern classification techniques, including (i) adaptive prototype learning algorithms that are particularly effective for data sets with a huge number of class types, (ii) randomized decomposition for support vector machine (SVM) that conducts SVM learning efficiently on large data sets based on a technique that enormously speeds up the training procedure and yet achieves relatively the same test accuracy using the existing technique, (3) feature selection method using multiple feature evaluation technique to select features for enhancing classification accuracy rates. These techniques can be applied to Chinese character recognition in which the character categories is huge (6,000 for commonly used characters and 130,000 for all characters), language classification for multi-linguistic documents, text categorization, ontology analysis, and many biological problems in which finding key features leads to improvement in classification accuracy rates.

 
 
bg