Wa Language Syllable Classification Using Support Multi-kernel Vector Machine Optimized by Immune Genetic Algorithm
| Autor: | Hua Yang, Huazhen Dong, Wenlin Pan, Meijun Fu |
|---|---|
| Rok vydání: | 2018 |
| Předmět: | |
| Zdroj: | Communications in Computer and Information Science ISBN: 9789811308956 GSKI (2) |
| DOI: | 10.1007/978-981-13-0896-3_51 |
| Popis: | A novel Wa syllable classification method based on multi-kernel support vector machine (MKSVM) optimized by immune genetic algorithm (IGA) is proposed in this paper. First, use vowel main body extension (VMBE) to extract the first dynamic characteristic parameter, pitch frequency. Then, use adaptive variational mode decomposition (AVMD) to extract the second dynamic characteristic parameter, formant frequency. Next, extract the mean values, standard errors, minima and maxima from the pitch frequency sequence and the first three formant frequency sequences respectively. Again, the feature sets with the mean values, standard errors, minima, maxima and label information, are inputted to IGA optimized MKSVM for analysis mode identification. Theoretical analysis demonstrates that MKSVM can approximate any multivariable function. The global optimal parameter vector of MKSVM can be rapidly identified by IGA parameter optimization. The experimental result of Wa syllable classification shows that, the proposed method significantly increases the accuracy of syllable classification and enhances the generalization of its application, and that, therefore, is feasible and effective on Wa language syllable classification. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|