Using data local information, the recently proposed local Fisher Discriminant Analysis (LFDA) algorithm (Sugiyama, 2007) provides a new way of handling the multimodal issues within classes where the conventional Fisher Discriminant Analysis(FDA) algorithm fails. Like the FDA algorithm ' its global counterpart FDA algorithm, the LFDA suffers when it is applied to the higher dimensional data sets.In this paper we propose a new formulation by which a robust algorithm can beformed. The new algorithm offers more robust results for higher dimensional datasets when compared with the LFDA in most cases. By extensive simulation studies,we have demonstrated the practical usefulness and robustness of our new algorithmin data visualization.
|Number of pages||15|
|Journal||International Journal of Pattern Recognition and Artificial Intelligence|
|Publication status||Published - Sep 2009|