Loading...
Please wait, while we are loading the content...
Similar Documents
The generalized risk zone and observations selection.
| Content Provider | CiteSeerX |
|---|---|
| Author | Peres, Rodrigo T. Carlos Pedreira, E. |
| Abstract | Abstract – We extend the risk zone concept by creating the Generalized Risk Zone. The Generalized Risk Zone is applied in a model-independent methodology to select representative observations in a sample set, with the goal of enhancing classification performance. The methodology involves the calculation of Cauchy-Schwartz divergence, as a measure of distance between densities, and we do so in the context of Information Theoretic Learning. We applied this methodology in Neural Networks, Support Vector Machines and Learning Vector Quantization. We have also discussed the comparison between Support Vectors and the vector that lay in the Generalized Risk Zone. |
| File Format | |
| Access Restriction | Open |
| Subject Keyword | Generalized Risk Zone Observation Selection Generalized Risk Zone Sample Set Support Vector Machine Information Theoretic Learning Classification Performance Model-independent Methodology Support Vector Risk Zone Concept Representative Observation Cauchy-schwartz Divergence Neural Network Learning Vector Quantization |
| Content Type | Text |
