改进的支持向量预选取方法在语音识别中的应用

系统仿真学报 ›› 2015, Vol. 27 ›› Issue (11): 2714-2721.

改进的支持向量预选取方法在语音识别中的应用

郝瑞¹, 牛砚波², 修磊³

1.山西财经大学信息管理学院,太原 030006;
2.太原理工大学信息工程学院,太原 030024;
3.山西财经大学统计学院,太原 030006

收稿日期:2014-12-24 修回日期:2015-03-30 出版日期:2015-11-08 发布日期:2020-08-05

Improved Support Vector Pre-extracting Algorithm in Speech Recognition Application

Hao Rui¹, Niu Yanbo², Xiu Lei³

1. College of Information Management, Shanxi University of Finance & Economics, Taiyuan 030006, China;
2. College of Information Engineering, Taiyuan University of Technology, Taiyuan 030024, China;
3. College of Statistics, Shanxi University of Finance & Economics, Taiyuan 030006, China

Received:2014-12-24 Revised:2015-03-30 Online:2015-11-08 Published:2020-08-05
About author:Hao Rui (1978-), W, PhD, Lecturer. Research interests: Artificial Intelligence.
Supported by:
Shanxi Scholarship Council of China (2009-28); Natural Science Foundation of Shanxi Province (2009011022-2)

摘要/Abstract

摘要： 对于大规模数据量的语音识别问题,支持向量机的训练成为一个难题。预选取支持向量是解决这一难题的方法之一。提出一种新的支持向量预选取算法.一方面对原数据集的每类数据分别进行核模糊C均值聚类,将所有的聚类中心作为每类数据的表征集;另一方面根据支持向量的几何分布含义并借鉴支持向量机的多类分类算法中一对一方法的思路提取原数据集的边界样本作为预选取支持向量进行训练和预测,并将该算法应用于嵌入式语音识别系统中,实验结果表明：该方法提高了语音识别系统的训练效率,降低了计算代价,同时保持了较高的识别率。

关键词: 支持向量, 多类分类, 核模糊C聚类, 样本预选取算法, 语音识别系统仿真

Abstract: Support vector machine (SVM) training is difficult for large-scale data set of speech recognition. A new SVM pre-extracting algorithm was proposed. On the one hand, kernel Fuzzy C-Means clustering was separately performed on each class of original data set. All the cluster centers were as a representative set of each class. On the other hand, according to the geometric distribution of support vectors and combined with the classification strategy of one-versus-one for SVM multi-class classification algorithm, boundary samples were extracted as support vectors for SVM to training and prediction. The algorithm was applied to embedded speech recognition system. Experiments indicate that this method improves the efficiency of training but also maintains the high recognition rate.

Key words: support vector, multi-class classification, kernel fuzzy C-Means clustering, sample pre- extracting, speech recognition system simulation

中图分类号:

TP391.9

郝瑞, 牛砚波, 修磊. 改进的支持向量预选取方法在语音识别中的应用[J]. 系统仿真学报, 2015, 27(11): 2714-2721.

Hao Rui, Niu Yanbo, Xiu Lei. Improved Support Vector Pre-extracting Algorithm in Speech Recognition Application[J]. Journal of System Simulation, 2015, 27(11): 2714-2721.

参考文献

[1] Hanilçi C, Ertaş F.Investigation of the effect of data duration and speaker gender on text-independent speaker recognition[J]. Computers & Electrical Engineering (S0045-7906), 2013, 39(2): 441-452.
[2] C H You, K A Lee, H Li.An SVM kernel with GMM-super vector based on the Bhattacharyya distance for speaker recognition[J]. IEEE Signal Processing Letters (S1070-9908), 2009, 16(1): 49-52.
[3] Sangeetha J, Jothilakshmi S.A novel spoken keyword spotting system using support vector machine[J]. Engineering Applications of Artificial in Intelligence (S0952-1976), 2014, 36(11): 287-293.
[4] Zhifeng Hao, Shu Yu, Xiaowei Yang, et al. Online LS-SVM Learning for Classification Problems Based on Incremental Chunk[C]// Lecture Notes in Computer Science (S0045-7906), 2004, 3173: 558-564.
[5] E Osuna, R Freund, F Girosi.An improved training algorithm machines [C]// Proceedings of the 1997 IEEE Workshop on Neural Net Processing. New York, USA: IEEE Press, 1997: 276-285.
[6] Palaniappan R, Sundaraj K, Sundaraj S.A comparative study of the svm and k-nn machine learning algorithms for the diagnosis of respiratory pathologies using pulmonary acoustic signals[J]. BMC Bioinformatics (S1471-2105), 2014, DOI: 10.1186/1471-2105-15-223.
[7] Jiao Licheng, Zhang Li, Zhou Weida.Pre-extracting Support Vectors for Support Vector Machine[J]. Electronica Sinica (S0372-2112), 2001, 29(3): 383-386.
[8] Zhang Bin, Tang Zhaohui, Zhu Hongqiu, et al. Novel Simplifying Method for Support Vector Machines and Its Application[J]. Journal of System Simulation (S1004-731X), 2012, 24(2): 344-247.
[9] Yang Jing, Yu Xu, Xie Zhiqiang.Support Vectors Pre-Extracting Method based on Improved Vector Projection[J]. Chinese Journal of Computers (S0254-4164), 2012, 35(5): 1002-1010.
[10] Ye Qiaolin, Ye Ning, Cui Jing, et al. Multisurface Support Vector Machines via Weight Vector Projection[J]. Pattern Recognition and Artificial Intelligence (S1003-6059), 2010, 23(5): 708-713.
[11] Zhang Jianpei, Zhao Ying, Yang Jing.Incremental Learning Algorithm of Support Vector Machine Based on Vector Projection[J]. Computer Science (S1002-137X), 2008, 35(3): 164-166.
[12] Arruti A, Mendialdua I, Sierra B, et al. New One Versus (One)(All) method: NOV@[J]. Expert Systems with Applications (S0957-4174), 2014, 41(10): 6251-6260.

[1]	李晖, 云昊, 岳红丽. 基于SVM逆系统的PMSM云模型整定PID控制[J]. 系统仿真学报, 2021, 33(8): 1846-1855.
[2]	位晶晶, 刘勤明, 叶春明, 李冠林. 基于GA-SVR的小样本数据缺失下的设备故障诊断[J]. 系统仿真学报, 2021, 33(6): 1342-1349.
[3]	邵诚, 孙宏远, 张林. 基于PSO-LSSVM的反舰导弹自控弹道仿真[J]. 系统仿真学报, 2021, 33(4): 918-926.
[4]	刘薛勤, 刘宁, 苏中, 王靖骁, 袁超杰. 基于足底压力感知的智能步态识别方法研究[J]. 系统仿真学报, 2021, 33(11): 2572-2578.
[5]	安玉, 焦朋朋, 白紫秀. 考虑多因素的驾驶行为安全评价与风险等级预测[J]. 系统仿真学报, 2021, 33(1): 118-126.
[6]	樊海霞, 陈小惠. 一种连续无创血压预测的改进向量机学习方法[J]. 系统仿真学报, 2020, 32(9): 1686-1692.
[7]	邢志伟, 姜松岳, 罗谦, 罗晓. 基于LWSVR的繁忙机场航班滑出时间预测[J]. 系统仿真学报, 2020, 32(5): 927-935.
[8]	於万里, 王艳, 纪志成. 氨糖发酵过程建模与工艺参数优化研究[J]. 系统仿真学报, 2020, 32(10): 1895-1902.
[9]	张宪霞, 章进强, 李致远, 马世伟, 杨帮华. 基于支持向量回归机学习的机械臂视觉反馈模糊控制[J]. 系统仿真学报, 2020, 32(10): 1997-2009.
[10]	张青, 邹湘军, 林桂潮, 孙艳辉. 草莓重量和形状图像特征提取与在线分级方法[J]. 系统仿真学报, 2019, 31(1): 7-9.
[11]	陈莉, 张海侠. 基于熵权-云模型的我国绿色智慧城市评价[J]. 系统仿真学报, 2019, 31(1): 136-144.
[12]	刘芳, 路丽霞, 王洪娟, 王鑫. 基于稀疏自动编码器和支持向量机的图像分类[J]. 系统仿真学报, 2018, 30(8): 3007-3014.
[13]	徐继亚, 王艳, 严大虎, 纪志成. 融合KPCA与信息粒化的滚动轴承性能退化SVM预测[J]. 系统仿真学报, 2018, 30(6): 2345-2354.
[14]	崔桂梅, 杨海靳, 刘丕亮, 于凯. 基于最小二乘支持向量机的氧化铝浓度预测[J]. 系统仿真学报, 2018, 30(5): 1844-1849.
[15]	王丽婕, 张利, 张岩. 基于CEEMD和混沌理论的超短期风电功率预测模型[J]. 系统仿真学报, 2018, 30(4): 1560-1565.