Linear components of quadratic classifiers
| Autor: | Javier Cárcamo, José R. Berrendero |
|---|---|
| Přispěvatelé: | UAM. Departamento de Matemáticas |
| Rok vydání: | 2018 |
| Předmět: |
Statistics and Probability
Computer science Oblique classification trees Matemáticas Quadratic discriminant analysis 01 natural sciences Piecewise linear function 010104 statistics & probability 03 medical and health sciences 0302 clinical medicine Quadratic equation 0101 mathematics Linear combination Reduction of the dimension business.industry Applied Mathematics Oblique case Pattern recognition Fisher linear discriminant analysis Quadratic classifier Computer Science Applications ComputingMethodologies_PATTERNRECOGNITION Hyperplane 030220 oncology & carcinogenesis Classification rule Supervised classification Feature extraction Artificial intelligence business Classifier (UML) |
| Zdroj: | Biblos-e Archivo. Repositorio Institucional de la UAM instname |
| Popis: | This is pre-print of an article published in Advances in Data Analysis and Classification. The final authenticated version is available online at: https://doi.org/10.1007/s11634-018-0321-6 We obtain a decomposition of any quadratic classifier in terms of products of hyperplanes. These hyperplanes can be viewed as relevant linear components of the quadratic rule (with respect to the underlying classification problem). As an application, we introduce the associated multidirectional classifier; a piecewise linear classification rule induced by the approximating products. Such a classifier is useful to determine linear combinations of the predictor variables with ability to discriminate. We also show that this classifier can be used as a tool to reduce the dimension of the data and helps identify the most important variables to classify new elements. Finally, we illustrate with a real data set the use of these linear components to construct oblique classification trees This research was supported by the Spanish MCyT grant MTM2016-78751-P |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|