Exploiting Global Input/output Access Pattern Classiication

Pattern Classi cation Using Hidden Markov Models

Input/output performance on current parallel le systems is sensitive to a good match of application access pattern to le system capabilities. Automatic input/output access classiication can determine application access patterns at execution time, guiding adaptive le system policies. In this paper we examine a new method for access pattern classiication that uses hidden Markov models, trained on...

Exploiting Global Input / Output Access Pattern Classi cation

Constructive Neural Network Learning Algorithms for Pattern Classiication

Constructive learning algorithms ooer an attractive approach for the incremental construction of near-minimal neural network architectures for pattern classiication. They help overcome the need for ad hoc and often inappropriate choices of network topology in the use of algorithms that search for suitable weights in a priori xed network architectures. Several such algorithms have been proposed ...

Modeling Fuzzy Classiication Systems with Compact Rule Base

An adaptive method to construct compact fuzzy systems for solving pattern classiication problems is presented. The method consists of two phases: a rule identiication phase and a rule selection phase. The rule identiication phase generates fuzzy rules from numerical data through a simple fuzzy grid method, then tunes the resulting fuzzy rules by training a neuro-fuzzy network used to model the ...

Novelty, Confidence & Errors in Connectionist Systems

Connectionist systems, or neural networks, have made a signiicant impact in the world of statistical pattern classiication, detection and prediction over the last few years. It is clear that the basic mechanisms involved with the method are sound. It is desirable, however, in all data analysis applications, to have estimates of the errors associated with the network's outputs. Whilst this is an...