本文選題：語音識(shí)別 + 最小二乘支持向量機(jī)　； 參考：《太原理工大學(xué)》2014年碩士論文

【摘要】：語音識(shí)別是一種最直接、最便捷的人機(jī)交互手段,屬于多維模式識(shí)別的范疇。最小二乘支持向量機(jī)是機(jī)器學(xué)習(xí)領(lǐng)域目前研究較熱的一種模式識(shí)別算法,作為標(biāo)準(zhǔn)支持向量機(jī)的一種擴(kuò)展,具有小樣本學(xué)習(xí)、能夠避免“高維維數(shù)災(zāi)難”和模型訓(xùn)練算法簡單易實(shí)現(xiàn)的優(yōu)點(diǎn),因此適用于復(fù)雜的語音信號(hào)的識(shí)別。但其具有解的稀疏性缺失的缺點(diǎn),造成模型復(fù)雜度的提高和系統(tǒng)識(shí)別速度的降低,本文針對(duì)這個(gè)問題展開了研究,具體研究內(nèi)容如下： (1)深入研究了語音識(shí)別系統(tǒng)和最小二乘支持向量機(jī)原理,將最小二乘支持向量機(jī)引入到語音識(shí)別系統(tǒng)中,克服了傳統(tǒng)語音識(shí)別方法中隱馬爾可夫模型需要先驗(yàn)分布知識(shí)、人工神經(jīng)網(wǎng)絡(luò)容易出現(xiàn)“過學(xué)習(xí)”的缺陷。 (2)仔細(xì)研究了模型參數(shù)對(duì)系統(tǒng)的學(xué)習(xí)能力和泛化能力的重要性,提出采用粒子群全局優(yōu)化算法結(jié)合K折交叉驗(yàn)證的方案進(jìn)行最佳參數(shù)尋優(yōu),避免了人工手動(dòng)調(diào)試復(fù)雜和網(wǎng)格算法耗時(shí)長的問題。 (3)在深入研究最小二乘支持向量機(jī)稀疏性缺失的原因和語音樣本特征維數(shù)對(duì)模型性能影響的基礎(chǔ)上,提出采用基于獨(dú)立成分分析的最小二乘支持向量機(jī)稀疏化方法。該方法首先采用獨(dú)立成分分析方法進(jìn)行語音特征降維；然后在模型訓(xùn)練之后采用基于獨(dú)立成分分析的快速剪枝算法對(duì)核矩陣進(jìn)行約簡,約簡過程中采用峰度和偏度的組合作為獨(dú)立成分重要性的度量指標(biāo),以此來解決獨(dú)立成分的排序問題。韓語語音庫上的實(shí)驗(yàn)表明,該算法在有效實(shí)現(xiàn)模型稀疏化的同時(shí)保證了模型識(shí)別精度。 (4)針對(duì)非支持向量參與模型訓(xùn)練會(huì)造成模型復(fù)雜度提高和模型識(shí)別性能降低的問題,本文從數(shù)據(jù)挖掘和支持向量的幾何分布含義兩個(gè)方面出發(fā),提出了基于支持向量預(yù)選取的最小二乘支持向量機(jī)稀疏化算法。該算法在模型訓(xùn)練之前,將K均值聚類算法提取的關(guān)鍵表征樣本和中心距離比值算法選取的邊界樣本的并集作為預(yù)選支持向量,從而有效實(shí)現(xiàn)了稀疏化。經(jīng)韓語語音庫和Aurora-2語音庫實(shí)驗(yàn)表明,該方法在幾乎不損失識(shí)別精度的基礎(chǔ)上提高了識(shí)別速度,達(dá)到了稀疏化的目的。
[Abstract]:Speech recognition is the most direct and convenient means of human-computer interaction, which belongs to the category of multidimensional pattern recognition. Least squares support vector machine (LS-SVM) is a hot pattern recognition algorithm in the field of machine learning. As an extension of standard SVM, LS-SVM has small sample learning. It can avoid the "high dimension disaster" and the advantages of simple and easy to implement the model training algorithm, so it is suitable for the recognition of complex speech signals. However, it has the disadvantage of lack of sparse solution, which leads to the increase of model complexity and the reduction of system recognition speed. This paper studies this problem, and the specific research contents are as follows: In this paper, the principle of speech recognition system and least square support vector machine (LS-SVM) is deeply studied, and the LS-SVM is introduced into speech recognition system, which overcomes the need of prior distribution knowledge in traditional speech recognition methods. Artificial neural network is prone to the defect of "overlearning". (2) the importance of model parameters to the learning ability and generalization ability of the system is studied carefully, and the particle swarm optimization algorithm combined with K-fold cross-validation is proposed to optimize the optimal parameters. The complex manual debugging and the time-consuming grid algorithm are avoided. 3) based on the in-depth study of the reasons for the lack of sparsity of LS-SVM and the effect of speech sample feature dimension on the performance of the model, an independent component analysis (ICA) based least-squares SVM thinning method is proposed. The method firstly uses independent component analysis (ICA) to reduce the dimension of speech features, and then, after model training, a fast pruning algorithm based on ICA is used to reduce the kernel matrix. The combination of kurtosis and skewness is used as a measure of the importance of independent components in the process of reduction, so as to solve the problem of sorting independent components. The experiments on the Korean language corpus show that the algorithm not only realizes the sparse model but also ensures the accuracy of model recognition. 4) aiming at the problem that non-support vector participation in model training will lead to higher model complexity and lower model recognition performance, this paper starts from two aspects: data mining and geometric distribution meaning of support vector. A least squares support vector machine thinning algorithm based on support vector preselection is proposed. Before the model training, the union of the key representation samples extracted by the K-means clustering algorithm and the boundary samples selected by the centroid distance ratio algorithm is taken as the pre-selected support vector. The experiments of Korean phonetic corpus and Aurora-2 corpus show that the method improves the recognition speed and achieves the purpose of thinning on the basis of almost no loss of recognition accuracy.
【學(xué)位授予單位】：太原理工大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2014
【分類號(hào)】：TN912.3

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 丁世飛;齊丙娟;譚紅艷;;支持向量機(jī)理論與算法研究綜述[J];電子科技大學(xué)學(xué)報(bào);2011年01期

2 陳亞秋,胡上序,陳德釗;前傳神經(jīng)網(wǎng)絡(luò)規(guī)模優(yōu)化的快速剪枝策略及其應(yīng)用[J];化工學(xué)報(bào);2001年06期

3 趙丹;馬勝前;鄭杰;;基于SPIHT編碼的語音信號(hào)壓縮算法[J];計(jì)算機(jī)工程與應(yīng)用;2011年09期

4 孔波;劉小茂;張鈞;;基于中心距離比值的增量支持向量機(jī)[J];計(jì)算機(jī)應(yīng)用;2006年06期

5 劉小茂;孔波;高俊斌;張鈞;;一種稀疏最小二乘支持向量分類機(jī)[J];模式識(shí)別與人工智能;2007年05期

6 樊繼聰;王友清;秦泗釗;;聯(lián)合指標(biāo)獨(dú)立成分分析在多變量過程故障診斷中的應(yīng)用[J];自動(dòng)化學(xué)報(bào);2013年05期

7 趙文杰;張立鶴;;基于約簡核矩陣的稀疏最小二乘支持向量機(jī)[J];計(jì)算機(jī)仿真;2013年07期

8 汪海燕;黎建輝;楊風(fēng)雷;;支持向量機(jī)理論及算法研究綜述[J];計(jì)算機(jī)應(yīng)用研究;2014年05期

9 梁錦錦;吳德;;稀疏L1范數(shù)最小二乘支持向量機(jī)[J];計(jì)算機(jī)工程與設(shè)計(jì);2014年01期

10 行鴻彥;金天力;;基于對(duì)偶約束最小二乘支持向量機(jī)的混沌海雜波背景中的微弱信號(hào)檢測[J];物理學(xué)報(bào);2010年01期

相關(guān)博士學(xué)位論文 前2條

1 呂釗;噪聲環(huán)境下的語音識(shí)別算法研究[D];安徽大學(xué);2011年

2 王法松;盲源分離的擴(kuò)展模型與算法研究[D];西安電子科技大學(xué);2013年

，

本文編號(hào)：1794692