發(fā)布時(shí)間：2018-04-18 11:22

  本文選題：流形學(xué)習(xí) + 低維可視化��； 參考：《江蘇師范大學(xué)》2014年碩士論文

【摘要】：漢語(yǔ)方言辨識(shí)研究的核心問(wèn)題之一是方言特征的提取,因?yàn)樘卣魈崛〉暮脡闹苯雨P(guān)系到系統(tǒng)性能的高低。傳統(tǒng)的特征提取大多數(shù)沿襲語(yǔ)種識(shí)別的理論和方法,忽視了漢語(yǔ)方言自身的有調(diào)性等區(qū)分意義較大的特征;其次,語(yǔ)音數(shù)據(jù)是一種典型的流形分布數(shù)據(jù),一些好的關(guān)于流形分布的算法沒(méi)有被應(yīng)用到漢語(yǔ)方言辨識(shí)中;傳統(tǒng)系統(tǒng)采用的單一特征在數(shù)據(jù)的結(jié)構(gòu)描寫(xiě)和信息的挖掘等方面有著很大的局限性。針對(duì)以上不足,論文將流形學(xué)習(xí)算法引入到漢語(yǔ)方言辨識(shí)系統(tǒng)中,從漢語(yǔ)方言語(yǔ)音低維可視化、流形學(xué)習(xí)算法提取漢語(yǔ)方言特征、流形學(xué)習(xí)算法的改進(jìn)和特征融合等幾個(gè)方面提升漢語(yǔ)方言辨識(shí)系統(tǒng)性能。具體成果如下:1.證明了漢語(yǔ)方言語(yǔ)音中流形結(jié)構(gòu)的存在。從低維可視化的角度對(duì)漢語(yǔ)方言語(yǔ)音進(jìn)行了分析研究,仿真實(shí)驗(yàn)結(jié)果表明相對(duì)于線性降維算法,流形學(xué)習(xí)算法在漢語(yǔ)方言語(yǔ)音低維的時(shí)候能夠更好地體現(xiàn)不同地區(qū)方言語(yǔ)音之間的差異性,間接證明了漢語(yǔ)方言語(yǔ)音數(shù)據(jù)中流形結(jié)構(gòu)的存在。2.利用流形學(xué)習(xí)提取了漢語(yǔ)方言辨識(shí)新特征。通過(guò)對(duì)低維流形結(jié)構(gòu)的觀察與流形學(xué)習(xí)算法的分析研究,利用其中局部線性嵌入算法對(duì)漢語(yǔ)方言語(yǔ)音進(jìn)行特征提取。3.在流形學(xué)習(xí)算法的基礎(chǔ)上對(duì)局部線性嵌入算法本身進(jìn)行了改進(jìn)。針對(duì)局部嵌入算法本身存在的不足,對(duì)其中的距離求取方法進(jìn)行改進(jìn),旨在改善樣本數(shù)據(jù)集的分布,并結(jié)合聚類(lèi)算法提取漢語(yǔ)方言語(yǔ)音新特征。4.構(gòu)建了一套基于流形學(xué)習(xí)的漢語(yǔ)方言辨識(shí)系統(tǒng),證明了新特征的有效性。利用高斯混合模型和支撐矢量機(jī)作為系統(tǒng)后端分類(lèi)器。仿真實(shí)驗(yàn)結(jié)果表明新特征可以有效地提高系統(tǒng)的性能。同時(shí)利用特征融合的方法對(duì)新特征和傳統(tǒng)特征進(jìn)行了有效地融合,進(jìn)一步提升了特征的有效性。
[Abstract]:The extraction of dialect features is one of the core problems in the research of Chinese dialect recognition, because the quality of feature extraction is directly related to the performance of the system.The traditional theories and methods of feature extraction mostly follow the language recognition, ignoring the tonality of Chinese dialects and other distinguishing features. Secondly, the speech data is a kind of typical manifold distribution data.Some good algorithms on manifold distribution have not been applied to Chinese dialect recognition, and the single feature used in traditional systems has great limitations in data structure description and information mining.In this paper, manifold learning algorithm is introduced into the Chinese dialect recognition system, and the feature of Chinese dialect is extracted from the low dimensional visualization of Chinese dialect speech and manifold learning algorithm.The improvement and feature fusion of manifold learning algorithm improve the performance of Chinese dialect recognition system.The concrete results are as follows: 1.The existence of manifold structure in Chinese dialect pronunciation is proved.In this paper, the Chinese dialect speech is analyzed and studied from the view of low dimension visualization. The simulation results show that compared with the linear dimensionality reduction algorithm,Manifold learning algorithm can better reflect the difference of dialect pronunciation in different regions when the dimension of Chinese dialect speech is low, which indirectly proves the existence of manifold structure in Chinese dialect phonetic data.The new features of Chinese dialect recognition are extracted by manifold learning.Based on the observation of low dimensional manifold structure and the analysis of manifold learning algorithm, the local linear embedding algorithm is used to extract the features of Chinese dialect speech.Based on the manifold learning algorithm, the local linear embedding algorithm is improved.In order to improve the distribution of sample data set and to extract the new features of Chinese dialect speech, the distance estimation method is improved to improve the distribution of the sample data set in order to overcome the shortcomings of the local embedding algorithm.A Chinese dialect recognition system based on manifold learning is constructed, which proves the validity of the new features.Gao Si hybrid model and support vector machine are used as the back-end classifier of the system.Simulation results show that the new features can effectively improve the performance of the system.At the same time, the method of feature fusion is used to fuse the new feature and the traditional feature effectively, which further improves the effectiveness of the feature.
【學(xué)位授予單位】：江蘇師范大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2014
【分類(lèi)號(hào)】：TN912.34

【相似文獻(xiàn)】

相關(guān)期刊論文 前10條

1 曾憲華;羅四維;;局部保持的流形學(xué)習(xí)算法對(duì)比研究[J];計(jì)算機(jī)工程與應(yīng)用;2008年29期

2 劉志勇;;基于保距與保拓?fù)涞牧餍螌W(xué)習(xí)算法[J];長(zhǎng)江大學(xué)學(xué)報(bào)(自然科學(xué)版)理工卷;2010年02期

3 閆志敏;劉希玉;;流形學(xué)習(xí)及其算法研究[J];計(jì)算機(jī)技術(shù)與發(fā)展;2011年05期

4 楊海紅;;流形學(xué)習(xí)中鄰域大小的選擇算法[J];山西煤炭管理干部學(xué)院學(xué)報(bào);2011年01期

5 周華;蔡超;丁明躍;;基于流形學(xué)習(xí)和流形高階近似的圖像距離度量[J];華中科技大學(xué)學(xué)報(bào)(自然科學(xué)版);2012年03期

6 石陸魁;張軍;宮曉騰;;基于鄰域保持的流形學(xué)習(xí)算法評(píng)價(jià)模型[J];計(jì)算機(jī)應(yīng)用;2012年09期

7 談超;關(guān)佶紅;周水庚;;增量與演化流形學(xué)習(xí)綜述[J];智能系統(tǒng)學(xué)報(bào);2012年05期

8 徐蓉;姜峰;姚鴻勛;;流形學(xué)習(xí)概述[J];智能系統(tǒng)學(xué)報(bào);2006年01期

9 羅四維;趙連偉;;基于譜圖理論的流形學(xué)習(xí)算法[J];計(jì)算機(jī)研究與發(fā)展;2006年07期

10 周紅;吳煒;滕奇志;楊曉敏;李e，

本文編號(hào)：1768187