In many machine learning problems, labeled training data is limited but unlabeled data is ample. Some of these problems have instances that can be factored into multiple views, ea...
We present two extensions of the algorithm by Broomhead et al [2] which is based on the idea that singular values that scale linearly with the radius of the data ball can be explo...
: Discovering interesting patterns or substructures in data streams is an important challenge in data mining. Clustering algorithms are very often applied to identify single substr...
The immune system provides an ideal metaphor for anomaly detection in general and computer security in particular. Based on this idea, artificial immune systems have been used for...
Language modeling is an effective and theoretically attractive probabilistic framework for text information retrieval. The basic idea of this approach is to estimate a language mo...