计算机工程与应用
計算機工程與應用
계산궤공정여응용
COMPUTER ENGINEERING AND APPLICATIONS
2010年
5期
136-139
,共4页
多约简支持向量机%模糊核聚类%说话人辨识%LDA变换
多約簡支持嚮量機%模糊覈聚類%說話人辨識%LDA變換
다약간지지향량궤%모호핵취류%설화인변식%LDA변환
Multi-Reduced Support Vector Machine(MRSVM)%kernel-based fuzzy clustering%speaker identification%LDA transform
提出一个新的基于MRSVM的说话人辨识方法,首先对语音特征矢量进行LDA降维,得到具有区分力的特征矢量,然后对其进行模糊核聚类,根据样本选择算法,选择聚类边界的特征矢量作为支持向量训练支持向量机,在不影响识别率的情况下,大大减少了支持向量杌的存储量和训练量.实验表明该方法具有较好的总体效果.
提齣一箇新的基于MRSVM的說話人辨識方法,首先對語音特徵矢量進行LDA降維,得到具有區分力的特徵矢量,然後對其進行模糊覈聚類,根據樣本選擇算法,選擇聚類邊界的特徵矢量作為支持嚮量訓練支持嚮量機,在不影響識彆率的情況下,大大減少瞭支持嚮量杌的存儲量和訓練量.實驗錶明該方法具有較好的總體效果.
제출일개신적기우MRSVM적설화인변식방법,수선대어음특정시량진행LDA강유,득도구유구분력적특정시량,연후대기진행모호핵취류,근거양본선택산법,선택취류변계적특정시량작위지지향량훈련지지향량궤,재불영향식별솔적정황하,대대감소료지지향량올적존저량화훈련량.실험표명해방법구유교호적총체효과.
A speaker identification method is proposed based on a novel Multi-Reduced Support Vector Machine(MRSVM).First ly,speech feature dimensions are reduced by using LDA transform;Secondly,the training data are selected at boundary of each cluster as Support Vectors(SVs)by using kemel-based fuzzy clustering technique.The experiment results show that the training data,time and storage can be reduced remarkably by using the proposed method without deteriorating recognition performance.The method is proved to be effective by the experiments.
