We have developed a methodology for predicting the performance of parallel algorithms on real parallel machines. The methodology consists of two steps. First, we characterize a mac...
Abstract—Gene expression data usually contains a large number of genes, but a small number of samples. Feature selection for gene expression data aims at finding a set of genes ...
We present an approach to the bid-evaluation problem in a system for multi-agent contract negotiation, called MAGNET. The MAGNET market infrastructure provides support for a variet...
Erik S. Steinmetz, John Collins, Scott Jamison, Ra...
In many text classification applications, it is appealing to take every document as a string of characters rather than a bag of words. Previous research studies in this area mostl...
In this paper, we evaluate and investigate two main types of relevance feedback algorithms; the Euclidean and the correlation?based approaches. In the first case, we examine heuri...