We propose a new stopping condition for a Support Vector Machine (SVM) solver which precisely reflects the objective of the Leave-OneOut error computation. The stopping condition ...
Promiscuous human leukocyte antigen (HLA) binding peptides are ideal targets for vaccine development. Existing computational models for prediction of promiscuous peptides used hidd...
By providing credit risk information, credit rating systems benefit most participants in financial markets, including issuers, investors, market regulators and intermediaries. In ...
The one-class and cost-sensitive support vector machines (SVMs) are state-of-the-art machine learning methods for estimating density level sets and solving weighted classificatio...
We present a random field based model for stereo vision with explicit occlusion labeling in a probabilistic framework. The model employs non-parametric cost functions that can be ...