An arti"cial neural network (ANN) algorithm is proposed that incorporates both market segmentation and discriminant (regression) analysis of the segments. The method simultan...
The advantage of a kernel method often depends critically on a proper choice of the kernel function. A promising approach is to learn the kernel from data automatically. In this p...
The research presented in this article is aimed at the development of an automated imaging system for classification of tissues in medical images obtained from Computed Tomography...
Background: Detecting remote homologies by direct comparison of protein sequences remains a challenging task. We had previously developed a similarity score between sequences, cal...
Abstract In case of insufficient data samples in highdimensional classification problems, sparse scatters of samples tend to have many ‘holes’—regions that have few or no nea...
Hakan Cevikalp, Diane Larlus, Marian Neamtu, Bill ...