面向语音识别的SVDD改进算法及仿真研究

doi:10.16182/j.issn1004731x.joss.201705011

系统仿真学报 ›› 2017, Vol. 29 ›› Issue (5): 1014-1020.doi: 10.16182/j.issn1004731x.joss.201705011

面向语音识别的SVDD改进算法及仿真研究

郝瑞¹, 刘晓峰², 牛砚波², 修磊¹

1.山西财经大学信息管理学院,太原 030006;
2.太原理工大学,太原 030024

收稿日期:2016-11-10 修回日期:2017-02-05 出版日期:2017-05-08 发布日期:2020-06-03

Improved SVDD for Speech Recognition and Simulation

Hao Rui¹, Liu Xiaofeng², Niu Yanbo², Xiu Lei¹

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

Received:2016-11-10 Revised:2017-02-05 Online:2017-05-08 Published:2020-06-03
About author:HAO Rui (1978-), W, PhD, Lecturer, Research interests: Artificial Intelligence.
Supported by:
National Natural Science Foundation of China (61072087), Shanxi University of Finance & Economics School youth fund(QN2015014)

摘要/Abstract

摘要： 支持向量数据描述(SVDD)将多类样本数据每一类用各自的超球来界定,显著降低了二次规划计算复杂度,更易于解决多类分类问题,因此在语音识别研究领域越来越受到广泛关注,本文针对语音样本分类中特征向量重叠和更新等问题,对现有的SVDD多类分类算法进行了改进,一方面,根据样本所在空间位置,构造超球重叠域决策函数;另一方面,基于类增量学习,实现超球类支持向量的动态改变。仿真实验结果表明,本文所提方法明显缩短了建模时间并且具有更好的识别性能。

关键词: 支持向量数据描述, 多类分类, 决策函数, 增量学习, 语音识别系统仿真

Abstract: Support vector data description (SVDD) defines multi-class data by their respective hyper-spheres. The computational complexity of the quadratic programming problem is reduced significantly and it is easier to solve multi-class classification problems. Thus, SVDD has attracted more and more attention in the field of speech recognition research. For the problems of the feature vectors of speech samples overlapping and updating, the conventional SVDD for multi-class classification was improved. On the one hand, the spatial position of the samples was fully used to construct the decision function in overlapping domain of hyper-spheres; On the other hand, based on class incremental learning the dynamic change of support vectors was implemented. Simulation experimental results indicate that the proposed method reduces modeling time obviously and has better recognition performance.

Key words: support vector data description, multi-class classification, decision function, incremental learning, speech recognition system simulation

中图分类号:

TP391.9

郝瑞, 刘晓峰, 牛砚波, 修磊. 面向语音识别的SVDD改进算法及仿真研究[J]. 系统仿真学报, 2017, 29(5): 1014-1020.

Hao Rui, Liu Xiaofeng, Niu Yanbo, Xiu Lei. Improved SVDD for Speech Recognition and Simulation[J]. Journal of System Simulation, 2017, 29(5): 1014-1020.

参考文献

[1] Chen Yanxiang, Xie Jian.Emotional speech recognition based on SVM with GMM[J]. Journal of Electronics (China)(S0372-2112), 2012, 29(3): 339-344.
[2] Zhang S X, Gales M J F. Structured SVMs for Automatic Speech Recognition[J]. IEEE Transactions on Audio Speech & Language Processing (S1558-7916), 2013, 21(3): 544-555.
[3] Elyes Zarrouk, Yassine Ben Ayed, Faiez Gargouri.Hybrid continuous speech recognition systems by HMM, MLP and SVM: a comparative study[J]. International Journal of Speech Technology (S1381-2416), 2014, 17(3): 223-233.
[4] Sangeetha J, Jothilakshmi S.Speech translation system for english to dravidian languages[J]. Applied Intelligence (S0924-669X), 2016, 46(3): 534-550.
[5] Kelly D, Caulfield B.Pervasive Sound Sensing: A Weakly Supervised Training Approach[J]. IEEE Transactions on Cybernetics (S2567-5471), 2016, 46(1): 123-135.
[6] Yang N, Yuan J, Zhou Y, et al.Enhanced multiclass SVM with thresholding fusion for speech-based emotion classification[J]. International Journal of Speech Technology(S1381-2416), 2016, 20(1): 27-41.
[7] Hao R, Niu Y B, Xiu L.Improved Support Vector Pre-Extracting Algorithm in Speech Recongnition Application[J]. Journal of System Simulation (S1004-731X), 2015, 27(11): 2714-2721.
[8] Ren X Y, Qi Y Z.Hadoop-based Multi-classiﬁcation Fusion for Intrusion Detection[J]. Journal of Applied Sciences (S1812-5654), 2013, 13(12): 2178-2181.
[9] Wang T, Zhao D, Feng Y.Two-stage multiple kernel learning with multiclass kernel polarization[J]. Knowledge-Based Systems (S0950-7051), 2013, 48(2): 10-16.
[10] Yi H, Song X, Jiang B.Structure selection for DAG-SVM based on misclassification cost minimization[J]. International Journal of Innovative Computing Information & Control (S1349-4198), 2011, 7(9): 5133-5143.
[11] Tax D M J,Duin R P W. Support Vector Data Description[J]. Machine Learning (S0885-6125), 2004, 54(1): 45-66.
[12] Zhou Y, Zhang X, Wang J, et al.Speaker recognition based on the combination of GMM and SVDD[J]. Przeglad Elektrotechniczny (S0033-2097), 2011, 87(3): 329-332.
[13] Fang J, Wang W, Wang X, et al.A SVDD Method Based on Maximum Distance between Two Centers of Spheres[J]. Chinese Journal of Electronics (S1022-4653), 2012, 21(1): 107-111.
[14] Lee J, Jin L, Park D, et al.Automatic Recognition of Aggressive Behavior in Pigs Using a Kinect Depth Sensor[J]. Sensors (S1605-0631), 2016, 16(5): 631.
[15] Wang Y, Yue J, Dong Y, et al.Review on Kernel based Target Tracking for Autonomous Driving[J]. Journal of Information Processing (S1882-6652), 2016, 24(1): 49-63.
[16] Gu B, Sheng V S, Wang Z, et al.Incremental learning for ν-Support Vector Regression[J]. Neural Networks the Official Journal of the International Neural Network Society (S0893-6080), 2015, 67(C): 140-150.
[17] Minervini M, Abdelsamea M M, Tsaftaris S A.Image-based plant phenotyping with incremental learning and active contours[J]. Ecological Informatics (S1574-9541), 2014, 23(23): 35-48.

面向语音识别的SVDD改进算法及仿真研究

Improved SVDD for Speech Recognition and Simulation

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 4

编辑推荐

Metrics

本文评价

[1]	刘三民, 刘余霞. 在线合成增量式数据流分类算法[J]. 系统仿真学报, 2018, 30(6): 2315-2320.
[2]	韩晓春, 薄翠梅, 易辉. 基于DNPE-SVDD的化工过程监控[J]. 系统仿真学报, 2018, 30(1): 184-190.
[3]	谢彦红, 孙呈敖, 李元. 加权动态SVDD在非线性过程监测中的应用研究[J]. 系统仿真学报, 2017, 29(7): 1506-1513.
[4]	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.