摘要
针对语音情感识别问题,提出一种采用决策模板的多分类器融合方法,利用不同类型的声学特征子集来构造子分类器。不同的子集能充分提高各子分类器之间的"多样性"指标,这是多分类器融合算法能够成功应用的必备条件。与多数投票融合算法和支持向量机相比该方法取得了较好的识别结果。另一方面,从多样性指标分析的角度出发探究该方法能获得较好识别效果的原因。
This paper proposes a novel scheme for speech emotion recognition,which uses Decision Templates(DT) ensemble algorithm to combine base classifiers built on acoustic feature subsets.Different feature subsets can provide sufficient diversity among base classifiers,which is known as a necessary condition for improvement in ensemble performance.Compared with those methods of Majority Voting ensemble and Support Vector Machine,the ensemble scheme proposed in this paper can achieve the highest performance.On the other hand,this paper investigates the possible reasons why ensemble systems can provide potential performance,in terms of diversity analysis.
出处
《计算机工程与应用》
CSCD
北大核心
2010年第28期205-207,235,共4页
Computer Engineering and Applications
作者简介
毕福昆(1982-),男,博士研究生,研究领域为模式识别;E—mail:bifukun@bit.edu.cn
边明明(1985-),男,博士研究生,研究领域为信号与信息处理,实时信号处理。