检索结果-内蒙古大学图书馆

Arithmetic Means of Accuracies: A Classifier Performance Measurement for Imbalanced Data Set

Arithmetic Means of Accuracies: A Classifier Performance Mea...

2010 International Conference on Audio,Language and Image Processing(2010年音频、语言与图像处理国际会议 ICALIP 2010)

作者： Ivanna K. Timotius Shaou-Gang Miaou Department of Electronic Engineering Satya Wacana Christian University Salatiga Indonesia Departm Department of Electronic Engineering Chung Yuan Christian University Chung Li Taiwan

ISBN: (纸本)9781424458561

Classifier performance measurement is essential in the development and analysis of classification algorithms. This paper proposes a new measurement approach that can be used generally for the balanced and imbalanced data set, can reflect the random guessing behavior perfectly, and can be used easily in cost-sensitive classification and multiple-class classification.

关键词： Dataset sorting algorithm Performance metrics arithmetical mean Classifiers ASSAY SCALE arithmetic Phyla Guess

来源：评论

学校读者我要写书评

暂无评论

Joint Gender-, Tone-, Vowel- Classification Via Novel Hierarchical Classification for Annotation of Monosyllabic Mandarin Word Tokens

Joint Gender-, Tone-, Vowel- Classification Via Novel Hierar...

引用

IEEE International Conference on Acoustics, Speech and Signal Processing

作者： Saurabh Garg Ghassan Hamarneh Allard Jongman Joan Sereno Yue Wang Pacific Parkinson's Research Centre University of British Columbia Canada Medical Image Analysis Lab Simon Fraser University (SFU) Canada KU Phonetics and Psycholinguistics Lab Department of Linguistics University of Kansas (KU) USA Language and Brain Lab Department of Linguistics Simon Fraser University Canada

ISBN: (纸本)9781538646595

The automatic annotation of Mandarin monosyllabic audio word tokens remains an important yet challenging issue in phonetics research. In this work, we address this annotation task via a novel subcategories-classification framework that not only performs word identification via the joint classifications of vowel and tone subcategories, but also performs gender discrimination of the speaker, which stands in contrast to previously proposed methods for Mandarin speech that focused only on tone-, vowel-, or gender- classification. We also propose a novel hierarchical classification algorithm to boost overall classification performance. Extensive experimental results show that our approach yielded superior performance in both cases of adequate and very limited training data. When trained using data from only one female and one male speaker, our approach also yielded the best classification accuracy in all subcategories of the token annotation problem, achieving an Fl-score of 0.742 as opposed to 0.705 as achieved by the second competing approach.

关键词： monosyllabic Mandarin tokens phonetics hierarchical classification label fusion multi-class S VMs phonetics Speech Sounds ANNOTATIONS Joints Loudspeakers Sex Discrimination token Taxonomy sorting algorithm

来源：评论

学校读者我要写书评

暂无评论

Comparison of Detection and Classification algorithms Using Boolean and Fuzzy Techniques

引用

ADVANCES IN FUZZY SYSTEMS 2012年第unknown期2012卷 406204.1-406204.10页

作者： Dixit, Rahul Singh, Harpreet Wayne State Univ Dept Engn & Comp Engn Detroit MI 48202 USA Wayne State Univ Dept Elect & Comp Engn Detroit MI 48202 USA

Modern military ranging, tracking, and classification systems are capable of generating large quantities of data. Conventional "brute-force" computational techniques, even with Moore's law for processors, present a prohibitive computational challenge, and often, the system either fails to "lock onto" a target of interest within the available duty cycle, or the data stream is simply discarded because the system runs out of processing power or time. In searching for high-fidelity convergence, researchers have experimented with various reduction techniques, often using logic diagrams to make inferences from related signal data. Conventional Boolean and fuzzy logic systems generate a very large number of rules, which often are difficult to handle due to limitations in the processors. Published research has shown that reasonable approximations of the target are preferred over incomplete computations. This paper gives a figure of merit for comparing various logic analysis methods and presents results for a hypothetical target classification scenario. Novel multiquantization Boolean approaches also reduce the complexity of these multivariate analyses, making it possible to better use the available data to approximate target classification. This paper shows how such preprocessing can reasonably preserve result confidence and compares the results between Boolean, multi-quantization Boolean, and fuzzy techniques.

关键词： Fuzzy multivariate analysis Quality factor PROCESSOR sorting algorithm Taxonomy processing power Targets logic diagram LARGE QUANTITIES data flow fuzzy logic system DUTY CYCLE

来源：评论

学校读者我要写书评

暂无评论

Detection of Deception Using Facial Expressions Based on Different Classification algorithms

Detection of Deception Using Facial Expressions Based on Dif...

引用

Scientific Conference of Electrical Engineering (SCEE)

作者： Harith H. Thannoon Wissam H. Ali Ivan A. Hashim Department of Electrical Engineering Nineveh University Mosul Iraq Department of Electrical Engineering University of Technology-Iraq Baghdad Iraq

ISBN: (纸本)9781728115887;9781728115870

Most psychologists argue that facial behavioral during lying is different from facial behavioral when telling the truth, so the facial behavioral can be used as reliable indicators for spotting liars. The deception detection systems (DDSs) based on facial expressions are non-invasive, mobile and cost-effective. In this work a DDS is dependent on Facial Action Coding System (FACS) for facial features extraction, the main idea of FACS is to describe all facial actions using Action Units (AUs), each AU is related to movement one or more facial muscles. Eight AUs are used which incorporated into a single facial behavior pattern vector; these AUs are AUs 5, 6, 7, 10, 12, 14, 23, and 28. Datasets are collected for 43 subjects (20males, 23 females) most of them between ages 18-25. Four types of classification algorithms are used individually in the last stage of the proposed system; these classifiers are MLP, KNN, VG-RAM, and SVM. The simulation results show that the best results are obtaining when using VG-RAM and KNN classifiers. The main contributions of this work are new classification techniques in DDS, collect real database that can use to measure the performance of any DDS based on facial expressions, and select suitable facial features.

关键词： face recognition feature extraction image classification image coding multilayer perceptrons muscle nearest neighbour methods object detection random-access storage support vector machines Facial Expression random access machine Multilayer perceptrons Feature extraction Object detection Deception image coding Support Vector Network Image classification direct digital synthesis sorting algorithm Facial Recognition nearest neighbor algorithm Face Face

来源：评论

学校读者我要写书评

暂无评论

A novel gender classification method based on MapReduce

A novel gender classification method based on MapReduce

引用

IEEE International Conference on Software Engineering and Service Sciences (ICSESS)

作者： Tong Cui Haifeng Zhao Science and Technology on Information Systems Engineering Laboratory Nanjing China

ISBN: (纸本)9781479983544

A novel parallelize gender recognition method with MapReduce is presented, which successfully comprise several machine leaning algorithms which are employed for gender recognition. The mass of face sample images are gathered and separated as train dataset and test dataset, and Local Binary Pattern (LBP) features are extracted when those sample sets are pre-processed and made ready for following operations. And Principle Component Analysis (PCA) is applied to train dataset to extract the most distinguishing features. Three classification algorithms: Support Vector Machine(SVM), k-Nearest Neighborhood (k-NN) and Adaboost are implemented and compared to determine the most suitable and successful algorithm for gender parallelize machine learning (GPML). To achieve the shortest execution time, we propose to apply GPML with MapReduce to avoid parallelizing above three algorithms while also improving their scalability to big datasets. The results show that this method reduces the training computational complexity significantly when the number of computing nodes increases while gaining better speedup rates and extending performance than those on parallelize Adaboost.

关键词： face recognition feature extraction image classification learning (artificial intelligence) parallel processing principal component analysis support vector machines Adaboost GPML LBP features MapReduce PCA SVM classification algorithm face sample images gender classification method gender parallelize machine learning k-nearest neighborhood local binary pattern feature extraction machine leaning algorithms parallelized gender recognition method principle component analysis support vector machine test dataset train dataset training computational complexity gender recognition parallel meta-learning Support Vector Network principal components analysis sorting algorithm Image classification PARALLEL PROCESSING (COMPUTERS) Facial Recognition Passive Cutaneous Anaphylaxis Feature extraction Dataset Gender ADABOOST execution time Classification Methods

来源：评论

学校读者我要写书评

暂无评论

Automated Cardiac Disease Diagnosis Using Support Vector Machine

Automated Cardiac Disease Diagnosis Using Support Vector Mac...

引用

International Conference on Circuits, Systems, Communication and Information Technology Applications (CSCITA)

作者： Samir S. Yadav Shivajirao M. Jadhav Rohit G. Bonde Shivani T. Chaudhari Dr. Babasaheb Ambedkar Technological University Raigad India

ISBN: (数字)9781728123394

ISBN: (纸本)9781728123400

This research is intended for the creation of an automatic cardiovascular disease detection system with pre-marked SPECT (Single photon emission computed tomography)data for training the program and for subsequently supplying it with unclassified SPECT data for classification as an indication of probable heart disease or not. We used a well-known and widely used classification algorithm named Support Vector Machines (SVM) for the purposes of this classification. In other terms, you instruct the machine; the system learns from the inputs you give. When the training is over, you can provide the system with undefined SPECT data and it is immediately identified by the system Clearly, the accuracy of this classification would depend on the precision of the pictures of the training system, the support vector machine parameters, the sophistication and many more variables that are discussed later. The project is of considerable significance in today's world, particularly if businesses offer a million-dollarreward to anyone capable of building automated systems detecting nearly every Star Wars (such as transcoder) disease.

关键词： aNN Support Vector Machines Linear Kernel SPECT Gaussian Kernel SVM Support Vector Network single photon emission computed tomography Heart Diseases Taxonomy Disease diagnosis Detection systems sorting algorithm Training Photon emission Business

来源：评论

学校读者我要写书评

暂无评论

Towards Using Subpattern Distributions in Social Network Analysis

Towards Using Subpattern Distributions in Social Network Ana...

引用

IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining

作者： Benjamin Cabrera University Duisburg-Essen

ISBN: (纸本)9781509028474

Determining the frequencies and the distribution of small subgraph patterns in a large input graph is an important part of many graph based mining tasks such as Frequent Subgraph Mining (FSM) and Motif Detection. Due to the exponential number of such graph patterns the interpretation of the mining results is mostly limited to finding unexpectedly frequent patterns, and in general identifying few particularly interesting patterns to then understand their function in the network. However, the full distribution of patterns itself encodes much more information about the underlying graph. Looking at this pattern distribution could be of particular interest in Social Network Analysis because social networks seem to have more random noise and less hard structural constraints compared to other types of graphs. This makes it unlikely to find meaning in only the most frequent patterns. In this paper we contribute in two ways to the usage of pattern distributions in networks analysis. First, we introduce two different sampling-based algorithms for computing the mentioned pattern distributions. Second we sketch two original ideas which can be used to harness the information in pattern distributions for gaining insights on global and local properties of Social Networks.

关键词： Social network services Image edge detection Approximation algorithms Databases Context Classification algorithms Clustering algorithms distribution (property) Social network services Edge detection sorting algorithm Clustering algorithms Social network analysis Approximation algorithms Network analysis Database Social Networks subschema Line graph

来源：评论

学校读者我要写书评

暂无评论

SSVEP-based BCI: a "Plug & Play" Approach

SSVEP-based BCI: a "Plug & Play" Approach

引用

Annual International Conference of the IEEE Engineering in Medicine and Biology Society

作者： Niccolò Mora Ilaria De Munari Paolo Ciampolini the Department of Information Engineering Unversity of Parma Parco Area delle Scienze 181/A 43124 Parma Italy

ISBN: (纸本)9781424492695

Brain-Computer Interface (BCI) can provide users with an alternative/augmentative interaction path, based on the interpretation of their brain activity. Steady State Visual Evoked Potentials (SSVEP) paradigm has many appealing features, aiming at implementing BCI-enabled communication-control applications. In this paper, we present a complete signal processing chain for a self-paced, SSVEP-based BCI. The proposed approach mostly focuses at reducing the user effort in dealing with BCI, featuring no need of user-specific calibration or training. In this paper, the classification algorithm is introduced and first validated on offline waveforms, aiming at improving classification accuracy and minimizing the false positive rate. Then, implementation of an online, self-paced SSVEP BCI is illustrated. The scheme refers to a four-way choice and exploits discrimination between intentional control states and no-control ones. Good performance is achieved, both in terms of true positive rate (>94%), as well as low false positive rate (0.26 min~(-1)), even in experiments carried out outside lab-controlled conditions.

关键词： BCI SSVEP scheme refers False-positives sorting algorithm brain activity plugs four-way junction Brain-Computer Interfaces Visual Evoked Potentials

来源：评论

学校读者我要写书评

暂无评论

Unsupervised Hyperspectral Image Classification algorithm By Integrating Spatial-Spectral Information

Unsupervised Hyperspectral Image Classification Algorithm By...

引用

International Conference on Audio, Language and Image Processing

作者： Belkacem Baassou Mingyi He Shaohui Mei Yifan Zhang Shaanxi Provincial Key Laboratory of Information Acquisition and Processing (IAP) School of Electronics and Information Northwestern Polytechnical University Shaanxi Provincial Key Laboratory of Information Acquisition and Processing (IAP)School of Electronics and Information Northwestern Polytechnical University

ISBN: (纸本)9781467301732

An integrated spatial-spectral information algorithm for hyper spectral image classification is proposed, which uses spatial pixel association (SPA)by exploiting spectral information divergence (SID) and spectral clustering to reduce regions number and improve classification accuracy. Moreover, a class boundary correction method is also developed to minimize the misclassified pixels at the edge of each class and to solve the problem of merged classes. Experiments with hyper spectral data demonstrate the effectiveness and advantages of the proposed frame work over some traditional methods in term of classification accuracy.

关键词： frame work Image classification hyperspectral imagery sorting algorithm pixel TRADITIONAL METHODS Sudden Infant Death

来源：评论

学校读者我要写书评

暂无评论

FEATURE SELECTION BY HIGH DIMENSIONAL MODEL REPRESENTATION AND ITS APPLICATION TO REMOTE SENSING

FEATURE SELECTION BY HIGH DIMENSIONAL MODEL REPRESENTATION A...

引用

IEEE International Geoscience and Remote Sensing Symposium

作者： Gulsen Taskin Kaya Huseyin Kaya Okan K. Ersoy Istanbul Technical Univ. Informatics Institute Purdue Univ. School of Electrical and Computer Eng.

ISBN: (纸本)9781467311601

In supervised classification, performance of a classification algorithm depends on several factors when dealing with a data set with a large number of features. One of them is the size of training data used in classification with respect to the number of features. Having less number of training data than the number of features can cause a reduction in classification performance with the addition of more features, called Hughes effect. A large number of features does not also guarantee higher classification accuracy as some of these features may be irrelevant for predicting the class labels. Additionally, the more features used during the training process, the higher is the computational time elapsed. Therefore, in terms of classification accuracy and less computational time, dimensionality reduction in feature size should be considered during classification when having the data with a large number of features. Two categories for reduction of dimensionality have been commonly encountered in the literature;feature extraction and feature selection. In feature extraction methods, new features are generated by taking linear combination of existing features. Fisher's linear discriminant analysis (LDA) is one of the most commonly used methods which aim to estimate a linear transformation to map data from a high dimensional space into a low-dimensional subspace that maintains unchanged separability. There are also many feature extraction algorithms developed and applied to hyperspectral images in the literature. Benediktsson et al. used decision boundary feature extraction method (DBFE) in classification of hyperspectral data with neural networks [1]. Discriminant analysis feature extraction (DAFE) was also another feature extraction method used for dimensionality reduction in classification. Kuo et al. proposed a new nonparametric feature extraction method (NWFE) based on scatter matrices for reduction of features of high-dimensional data[2].

关键词： Feature extraction dimensionality reduction remote sensing sorting algorithm computational time Linear Combination linear transformations

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：