A Highly Accurate Dysphonia Detection System Using Linear Discriminant Analysis

The recognition of pathological voice is considered a difficult task for speech analysis.Moreover,otolaryngologists needed to rely on oral communication with patients to discover traces of voice pathologies like dysphonia that are caused by voice alteration of vocal folds and their accuracy is between 60%–70%.To enhance detection accuracy and reduce processing speed of dysphonia detection,a novel approach is proposed in this paper.We have leveraged Linear Discriminant Analysis(LDA)to train multiple Machine Learning(ML)models for dysphonia detection.Several ML models are utilized like Support Vector Machine(SVM),Logistic Regression,and K-nearest neighbor(K-NN)to predict the voice pathologies based on features like Mel-Frequency Cepstral Coefficients(MFCC),Fundamental Frequency(F0),Shimmer(%),Jitter(%),and Harmonic to Noise Ratio(HNR).The experiments were performed using Saarbrucken Voice Data-base(SVD)and a privately collected dataset.The K-fold cross-validation approach was incorporated to increase the robustness and stability of the ML models.According to the experimental results,our proposed approach has a 70%increase in processing speed over Principal Component Analysis(PCA)and performs remarkably well with a recognition accuracy of 95.24%on the SVD dataset surpassing the previous best accuracy of 82.37%.In the case of the private dataset,our proposed method achieved an accuracy rate of 93.37%.It can be an effective non-invasive method to detect dysphonia.

A Comparison of Two Linear Discriminant Analysis Methods That Use Block Monotone Missing Training Data

We revisit a comparison of two discriminant analysis procedures, namely the linear combination classifier of Chung and Han (2000) and the maximum likelihood estimation substitution classifier for the problem of classifying unlabeled multivariate normal observations with equal covariance matrices into one of two classes. Both classes have matching block monotone missing training data. Here, we demonstrate that for intra-class covariance structures with at least small correlation among the variables with missing data and the variables without block missing data, the maximum likelihood estimation substitution classifier outperforms the Chung and Han (2000) classifier regardless of the percent of missing observations. Specifically, we examine the differences in the estimated expected error rates for these classifiers using a Monte Carlo simulation, and we compare the two classifiers using two real data sets with monotone missing data via parametric bootstrap simulations. Our results contradict the conclusions of Chung and Han (2000) that their linear combination classifier is superior to the MLE classifier for block monotone missing multivariate normal data.

Direct linear discriminant analysis based on column pivoting QR decomposition and economic SVD

A direct linear discriminant analysis algorithm based on economic singular value decomposition （DLDA/ESVD） is proposed to address the computationally complex problem of the conventional DLDA algorithm, which directly uses ESVD to reduce dimension and extract eigenvectors corresponding to nonzero eigenvalues. Then a DLDA algorithm based on column pivoting orthogonal triangular （QR） decomposition and ESVD （DLDA/QR-ESVD） is proposed to improve the performance of the DLDA/ESVD algorithm by processing a high-dimensional low rank matrix, which uses column pivoting QR decomposition to reduce dimension and ESVD to extract eigenvectors corresponding to nonzero eigenvalues. The experimental results on ORL, FERET and YALE face databases show that the proposed two algorithms can achieve almost the same performance and outperform the conventional DLDA algorithm in terms of computational complexity and training time. In addition, the experimental results on random data matrices show that the DLDA/QR-ESVD algorithm achieves better performance than the DLDA/ESVD algorithm by processing high-dimensional low rank matrices.

Two linear subpattern dimensionality reduction algorithms

This paper presents two novel algorithms for feature extraction-Subpattern Complete Two Dimensional Linear Discriminant Principal Component Analysis (SpC2DLDPCA) and Subpattern Complete Two Dimensional Locality Preserving Principal Component Analysis (SpC2DLPPCA). The modified SpC2DLDPCA and SpC2DLPPCA algorithm over their non-subpattern version and Subpattern Complete Two Dimensional Principal Component Analysis (SpC2DPCA) methods benefit greatly in the following four points: (1) SpC2DLDPCA and SpC2DLPPCA can avoid the failure that the larger dimension matrix may bring about more consuming time on computing their eigenvalues and eigenvectors. (2) SpC2DLDPCA and SpC2DLPPCA can extract local information to implement recognition. (3)The idea of subblock is introduced into Two Dimensional Principal Component Analysis (2DPCA) and Two Dimensional Linear Discriminant Analysis (2DLDA). SpC2DLDPCA combines a discriminant analysis and a compression technique with low energy loss. (4) The idea is also introduced into 2DPCA and Two Dimensional Locality Preserving projections (2DLPP), so SpC2DLPPCA can preserve local neighbor graph structure and compact feature expressions. Finally, the experiments on the CASIA(B) gait database show that SpC2DLDPCA and SpC2DLPPCA have higher recognition accuracies than their non-subpattern versions and SpC2DPCA.

An Optimization Criterion for Generalized Marginal Fisher Analysis on Undersampled Problems

Marginal Fisher analysis （MFA） not only aims to maintain the original relations of neighboring data points of the same class but also wants to keep away neighboring data points of the different classes.MFA can effectively overcome the limitation of linear discriminant analysis （LDA） due to data distribution assumption and available projection directions.However,MFA confronts the undersampled problems.Generalized marginal Fisher analysis （GMFA） based on a new optimization criterion is presented,which is applicable to the undersampled problems.The solutions to the proposed criterion for GMFA are derived,which can be characterized in a closed form.Among the solutions,two specific algorithms,namely,normal MFA （NMFA） and orthogonal MFA （OMFA）,are studied,and the methods to implement NMFA and OMFA are proposed.A comparative study on the undersampled problem of face recognition is conducted to evaluate NMFA and OMFA in terms of classification accuracy,which demonstrates the effectiveness of the proposed algorithms.

DIMENSIONALITY REDUCTION BASED ON SVM AND LDA,AND ITS APPLICATION TO CLASSIFICATION TECHNIQUE

Some dimensionality reduction （DR） approaches based on support vector machine （SVM） are proposed. But the acquirement of the projection matrix in these approaches only considers the between-class margin based on SVM while ignoring the within-class information in data. This paper presents a new DR approach, call- ed the dimensionality reduction based on SVM and LDA （DRSL）. DRSL considers the between-class margins from SVM and LDA, and the within-class compactness from LDA to obtain the projection matrix. As a result, DRSL can realize the combination of the between-class and within-class information and fit the between-class and within-class structures in data. Hence, the obtained projection matrix increases the generalization ability of subsequent classification techniques. Experiments applied to classification techniques show the effectiveness of the proposed method.

基于子空间学习的快速自适应局部比值和判别分析

降维是处理高维数据的一项关键技术,其中线性判别分析及其变体算法均为有效的监督算法。然而大多数判别分析算法存在以下缺点:a)无法选择更具判别性的特征;b)忽略原始空间中噪声和冗余特征的干扰;c)更新邻接图的计算复杂度高。为了克服以上缺点,提出了基于子空间学习的快速自适应局部比值和判别分析算法。首先,提出了统一比值和准则及子空间学习的模型,以在子空间中探索数据的潜在结构,选择出更具判别信息的特征,避免受原始空间中噪声的影响;其次,采用基于锚点的策略构造邻接图来表征数据的局部结构,加速邻接图学习;然后,引入香农熵正则化,以避免平凡解;最后,在多个数据集上进行了对比实验,验证了算法的有效性。

改进LPCDA算法及其在旋转机械故障诊断中的应用

针对高维故障数据集中有效信息利用率低导致故障分类难度偏大的问题,提出一种线性主成分判别分析(linear principal component discriminant analysis,简称LPCDA)的故障数据集降维算法。该算法将类间可分性判别与主成分计算的思想融入线性判别分析(linear discriminant analysis,简称LDA)算法中,使算法拥有剔除相关信息和冗余特征的能力,从而可以更好地保留能够反映机械运行状态有价值的故障状态信息以及特征的主要成分。实验结果表明,本算法能够剔除高维故障数据集中的相关信息、冗余特征并保留特征主要成分,具有降低故障分类难度与提高自动辨识准确率的功能。该研究可为有效降低高维故障数据集的规模和故障的分类难度、提高有效信息的挖掘能力,提供了理论参考依据。

基于组态分析的数字直播活动效果提升路径研究

[目的/意义]数字直播是企业开展网络活动的重要方式。通过发现能够使企业数字直播活动创造更大价值的要素或要素组合,并采取多途径增强这些要素,从而达到企业数字直播活动效果提升的目的,这对企业发展有创新指导意义。[方法/过程]选取4个与数字直播活动相关的影响因素作为条件变量,采用活动效果作为结果变量,并结合清晰集定性比较分析与线性判别降维对烟草领域数字直播活动的数据进行组态分析。[结果/结论]研究发现,单个因素并不构成提升数字直播活动效果的必要条件,应考虑不同条件因素的组合;存在游戏辅助型和话题支撑型两条组态路径能改善数字直播活动的效果;这两条路径适用不同的数字直播形式,企业应根据自身需求进行选择并在实践中细化路径。

自适应近邻局部比值和线性判别分析算法

在机器学习和模式识别中,降维能够显著提升分类器的判别性能与效率。比率和(ratio sum,RS)是线性判别分析(linear discriminant analysis,LDA)的一种全新变体,它试图使投影矩阵在每个维度上都达到最优。但RS并没有考虑到数据的局部几何结构,这就可能导致无法求得最优解。为了克服RS的这一缺点,提出了一种自适应近邻局部比值和线性判别分析算法(adaptive neighbor local ratio sum linear discriminant analysis,ANLRSLDA)。该算法使用自适应近邻的构图方法构建邻接矩阵,保留数据的局部几何结构完成了数据类间及类内矩阵的构建,从而更好地找到数据的最优表示;并且该方法采用有效的无核参数邻域分配策略来构造邻接矩阵,避免调整热核参数的需要。在UCI数据集及人脸数据集进行了对比实验,验证了该算法的有效性。

基于局部边缘判别投影的发动机故障诊断方法

在线性判别分析(Linear Discriminant Analysis,LDA)的基础上,局部边缘判别投影(Locality Margin Discriminant Projection,LMDP)重新定义类间散布矩阵和类内散布矩阵,使得数据样本中异类样本在低维空间中的距离更远、同类样本在低维空间中的距离更近,增强数据样本的可区分度。为更好提取发动机的故障特征,实现发动机故障有效诊断,以LMDP为核心,结合特征提取方法和模式识别方法,给出基于LMDP的发动机故障诊断流程。发动机故障诊断结果表明,LMDP可实现发动机不同故障类型的有效区分,显著提升后续的诊断精度,具有一定的优势。

间隔影响分析波长选择算法在近红外光谱鉴别贝类毒素中的应用

该文采用近红外光谱技术与化学计量学方法结合实现贝类毒素无损鉴别。该研究以新鲜翡翠贻贝为研究对象,使用近红外光谱仪采集健康贻贝和感染腹泻性毒素贻贝的反射光谱数据,利用Savitzky-Golay卷积平滑求导结合标准正态变量变换光谱预处理方式消除光谱中的干扰因素,采用间隔影响分析(margin influence analysis,MIA)结合连续投影算法(successive projections algorithm,SPA)对数据进行降维处理,应用偏最小二乘线性判别分析(partial least squares linear discriminant analysis,PLS-LDA)方法构建贝类毒素鉴别模型,并与支持向量机和随机森林分析模型进行比较。结果表明,采用MIA-SPA-PLS-LDA方法,可实现贝类毒素的100%鉴别。为此,利用MIA-SPA-PLS-LDA方法可建立准确的贝类毒素鉴别模型,为贝类毒素的快速鉴别提供了新途径,也为后续各种贝类水产品的毒素鉴别分析提供了参考。

Fisher线性鉴别分析的理论研究及其应用

Fisher线性鉴别分析已成为特征抽取的最为有效的方法之一 .但是在高维、小样本情况下如何抽取Fisher最优鉴别特征仍是一个困难的、至今没有彻底解决的问题 .文中引入压缩映射和同构映射的思想 ,从理论上巧妙地解决了高维、奇异情况下最优鉴别矢量集的求解问题 ,而且该方法求解最优鉴别矢量集的全过程只需要在一个低维的变换空间内进行 ,这与传统方法相比极大地降低了计算量 .在此理论基础上 ,进一步为高维、小样本情况下的最优鉴别分析方法建立了一个通用的算法框架 ,即先作K L变换 ,再用Fisher鉴别变换作二次特征抽取 .基于该算法框架 ,提出了组合线性鉴别法 ,该方法综合利用了F S鉴别和J Y鉴别的优点 ,同时消除了二者的弱点 .在ORL标准人脸库上的试验表明 ,组合鉴别法所抽取的特征在普通的最小距离分类器和最近邻分类器下均达到 97%的正确识别率 ,而且识别结果十分稳定 .

基于双向二维最大间距准则的局部放电灰度图像特征提取

针对高压交联聚乙烯电力电缆中间接头绝缘缺陷的辨识问题,提出一种局部放电灰度图像特征提取的双向二维最大间距准则方法,对获取的局部放电灰度图像从水平和垂直2个方向进行投影,得到了不同类别灰度图的鉴别矢量,选用最近邻分类器进行局部放电分类,以辨识电缆中间接头出现的不同绝缘缺陷。该方法解决了局部放电灰度图像特征提取维数大、识别样本少的难题。在对实验室4种典型电缆接头绝缘缺陷产生的PD信号进行对比辨识表明,其局部放电特征提取的速度和绝缘缺陷的识别率优于常用的主成分分析或Fisher鉴别分析方法。

基于最大散度差鉴别准则的自适应分类算法

首先证明了,当类内散布矩阵非奇异时,特定参数值c_0下最大散度差的最优鉴别方向等同于Fisher最优鉴别方向；其次,给出了最大散度差分类算法的识别率随参数C变化的曲线．该曲线通常为一脉冲曲线．随着参数C的增大,识别率也逐渐增大．当参数C增大到c_0时,识别率达到最大值．另外,以往的研究成果表明：当类内散布矩阵奇异时,最大散度差鉴别准则逐步逼近大间距线性投影准则．而且,随着参数C的不断增大,最大散度差分类算法的识别率也单调增大并最终稳定到大间距线性投影分类算法的识别率上．为此,我们提出了基于最大散度差鉴别准则的自适应分类算法．新算法可以根据训练样本的特性(类内散布矩阵是否奇异)自动选择恰当的参数C．在UCI机器学习数据库上的6个数据集以及AR人脸图像数据库上的测试结果表明,自适应最大散度差分类算法具有良好的分类性能．

基于模块2DPCA的人脸识别方法

提出了模块2DPCA(two-d im ensional princ ipal component analysis)的人脸识别方法。模块2DPCA方法先对图像矩阵进行分块,将分块得到的子图像矩阵直接用于构造总体散布矩阵,然后利用总体散布矩阵的特征向量进行图像特征抽取。与基于图像向量的鉴别方法(比如PCA)相比,该方法在特征抽取之前不需要将子图像矩阵转化为图像向量,能快速地降低鉴别特征的维数,可以完全避免使用矩阵的奇异值分解,特征抽取方便;此外,模块2DPCA是2DPCA的推广。在ORL和NUST603人脸库上的试验结果表明,模块2DPCA方法在识别性能上优于PCA,比2DPCA更具有鲁棒性。

基于最大差值的二维边界Fisher的人脸识别

提出了一种基于最大差值的二维边界Fisher的鉴别分析方法。该方法利用描述类间数据可分性的相似度矩阵Sp与描述类内数据紧致性的相似度矩阵Sc之差作为鉴别准则,从而避免了边界Fisher鉴别分析所遇到的小样本问题。所提方法是直接基于图像矩阵的,与以往的基于图像向量的方法相比,进一步提高了识别的正确率。另外,还揭示了基于最大差值的边界Fisher鉴别方法和边界Fisher鉴别的内在关系。在ORL和Yale人脸数据库上的实验表明,所提方法具有较高的识别率。

一种适用于小样本问题的基于边界的特征提取算法

特征提取技术是模式识别领域进行数据降维和强化判别信息的有效方法．线性判别分析是监督特征提取方法的典型代表,获得广泛应用,但受到小样本问题的制约．对此提出一种适用于小样本问题的基于边界的特征提取算法．算法利用高维数据小样本情况下线性可分概率增加以及其低维投影趋于正态分布的特点,定义了新的类别边界,不但考虑了由线性判别分析提出的类内、类间离散度,也兼顾各类别的方差差异性．通过极大化该边界获得最优投影向量,同时避免因类内离散度矩阵奇异导致的小样本问题．进一步将算法推广到多类问题．高光谱数据特征提取与分类实验表明,算法在小样本情况下对于两类和多类问题均具有良好的推广性能,优于多种线性判别分析的改进算法,并且在样本较多时也取得了满意结果．

人脸与虹膜特征层融合模型的研究

多生物特征的融合与识别可提高身份识别系统的整体性能.本文在研究特征层融合的基础上,结合二维Fisher线性判别分析(2-Dimensional Fisher Linear Discriminant Analysis,2DFLD),提出了一种人脸与虹膜特征融合与识别模型.首先,对人脸图像与虹膜图像分别进行压缩降维处理,得到相应的初始特征矩阵.然后将人脸与虹膜的初始特征矩阵进行组合,获得组合特征矩阵.同时,利用2DFLD算法对组合特征矩阵进行融合,获得了人脸与虹膜的融合特征.最后运用最小距离分类器进行识别.基于ORL(Olivetti Research Laboratory)人脸数据库和CASIA(Chinese Academy ofSciences,Institute of Automation)虹膜数据库的实验结果表明,该模型实现了特征层融合,不仅克服了"小样本"效应,而且有效提高了身份识别的正确识别率,为多生物特征身份识别提供了一种有效模型.

二维主成分分析方法的推广及其在人脸识别中的应用

提出了分块二维主成分分析(分块2DPCA)的人脸识别方法。分块2DPCA方法先对图像矩阵进行分块,对分块得到的子图像矩阵直接进行鉴别分析。其特点是:能方便地降低鉴别特征的维数;可以完全避免使用矩阵的奇异值分解,特征抽取方便;与2DPCA方法相比,使用低维的鉴别特征矩阵,而达到较高(至少是不低)的正确识别率。此外,2DPCA是分块2DPCA的特例。在ORL和NUST603人脸库上的试验结果表明,所提出的方法在识别性能上优于2DPCA方法。