A Multi-SVM Classification System
| Autor: | Andreas Stafylopatis, Dimitrios S. Frossyniotis |
|---|---|
| Rok vydání: | 2001 |
| Předmět: |
Computer Science::Machine Learning
Scheme (programming language) Computer science business.industry Probabilistic logic Space (commercial competition) Machine learning computer.software_genre Linear subspace Fuzzy logic Support vector machine Statistics::Machine Learning ComputingMethodologies_PATTERNRECOGNITION Unsupervised learning Artificial intelligence Space partitioning business computer computer.programming_language |
| Zdroj: | Multiple Classifier Systems ISBN: 9783540422846 Multiple Classifier Systems |
| DOI: | 10.1007/3-540-48219-9_20 |
| Popis: | It has been shown by several researchers that multiclassifier systems can result in effective solutions to difficult tasks. In this work, we propose a multi-classifier system based on both supervised and unsupervised learning. According to the principle of "divide-and-conquer", the input space is partitioned into overlapping subspaces and Support Vector Machines (SVMs) are subsequently used to solve the respective classification subtasks. Finally, the decisions of the individual SVMs are appropriately combined to obtain the final classification decision. We used the Fuzzy c-means (FCM) method for input space partitioning and we considered a scheme for combining the decisions of the SVMs based on a probabilistic interpretation. Compared to single SVMs, the multi-SVM classification system exhibits promising accuracy performance on well-known data sets. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|