A Supervised Learning Technique and Its Applications to Computational Biology

The problem of classifying data in spaces with thousands of dimensions have recently been addressed in literature for its importance in computational biology. An example of such applications is the analysis of genomic and proteomic data. Among the most promising techniques that classify such data in lower dimensional subspace, Top Scoring Pairs has the advantage of finding a two-dimensional subspace with a simple decision rule. In the present paper we show how this technique can take advantage from the utilization of incremental generalized eigenvalue classifier to obtain higher classification accuracy with a small training set.