On the Optimization of Multiclass Support Vector Machines Dedicated to Speech Recognition
| Autor: | Assia Benyettou, Freha Mezzoudj |
|---|---|
| Rok vydání: | 2012 |
| Předmět: |
Structured support vector machine
Computer science Generalization business.industry Process (engineering) Speech corpus Feature selection Machine learning computer.software_genre Support vector machine Multiclass classification Genetic algorithm Instance selection Artificial intelligence business computer |
| Zdroj: | Neural Information Processing ISBN: 9783642344800 ICONIP (2) |
| DOI: | 10.1007/978-3-642-34481-7_1 |
| Popis: | We present in this paper an interesting approach to enhance the performance of multi-classification using Genetic Algorithm. Two systems for an instance selection and feature selection are respectively introduced. We combined Genetic Algorithm with multiclass Support Vector Machines in order to reduce the learning set. The goal is to simplify the learning process and to improve the generalization. The results obtained on speech corpus show encouraging improvements in terms of processing time and classification accuracies. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|