Feature extraction for classification problems and its application to face recognition

This study investigates a new method of feature extraction for classification problems. The method is based on the independent component analysis (ICA). However, unlike the original ICA, one of the unsupervised learning methods, it is developed for classification problems by utilizing class information. The proposed method is an extension of our previous work on binary-class problems to multi-class problems. It treats the class labels as input features in order to produce two sets of new features: one that carries much information on the class labels and the other that is irrelevant to the class. The learning rule for this method is obtained using the stochastic gradient method to maximize the likelihood of the observed data. Among the new features, using only class-relevant ones, the dimension of the feature space can be greatly reduced in line with the principle of parsimony, resulting better generalization. This method was applied to recognize face identities and facial expressions using various databases such as the Yale, AT&T (former ORL), Color FERET face databases and so on. The performance of the proposed method was compared with those of conventional methods such as the principal component analysis (PCA), Fisher's linear discriminant (FLD), etc. The experimental results show that the proposed method performs well for face recognition problems.     

A novel distance measure on q-rung picture fuzzy sets and its application to decision making and classification problems

Artificial Intelligence Review - In recent years, different higher order fuzzy sets have been introduced to better handle the uncertainty in many practical decision making and data mining problems....     

A fast classification scheme and its application to face recognition

To overcome the high computational complexity in real-time classifier design, we propose a fast classification scheme. A new measure called ’reconstruction proportion’ is exploited to reflect the discriminant information. A novel space called the ’reconstruction space’ is constructed according to the reconstruction proportions. A point in the reconstruction space denotes the case of a sample reconstructed using training samples. This is used to search for an optimal mapping from the conventional sample space to the reconstruction space. When the projection from the sample space to the reconstruction space is obtained, a new sample after mapping to the new discriminant space would be classified quickly according to the reconstruction proportions in the reconstruction space. This projection technique results in a diversion of time-consuming calculations from the classification stage to the training stage. Though training time is prolonged, it is advantageous in that classification problems such as identification can be solved in real time. Experimental results on the ORL, Yale, YaleB, and CMU PIE face databases showed that the proposed fast classification scheme greatly outperforms conventional classifiers in classification accuracy and efficiency.     

A novel similarity/dissimilarity measure for intuitionistic fuzzy sets and its application in pattern recognition

Among the most interesting measures in intuitionistic fuzzy sets (IFSs) theory, the similarity measure is an essential tool to compare and determine degree of similarity between IFSs. Although there exist many similarity measures for IFSs, most of them cannot satisfy the axioms of similarity measure or provide reasonable results. In this paper, a novel knowledge-based similarity/dissimilarity measure between IFSs is proposed. Firstly, we define a new knowledge measure of information conveyed by the IFS and prove some properties of the proposed knowledge measure. Based on the proposed knowledge measure of IFSs, we construct a novel similarity/dissimilarity measure between IFSs and prove some properties of the proposed similarity measure. Then we use some illustrative examples to show that the proposed measures, though simple in concept and calculus, can overcome the drawbacks of the existing measures. Finally, we apply the proposed similarity/dissimilarity measure between IFSs in the pattern recognition problems to demonstrate that the proposed measure is the most reliable to deal with the pattern recognition problem in comparison with the existing similarity measures.     

Optimal metrics in classification problems with ordered features and an arbitrary number of classes

This paper presents a method of searching for the best distance function in a classification problem with k classes in which the objects are represented by vectors of ordered features. An optimal distance function is sought with a linear criterion that minimizes the weighted difference of the average intra- and interclass distances. In the first part of the paper each feature space has its own distance function specified on the Cartesian product of integer numbers from 0 to N − 1 with its value in an interval from 0 to M − 1. The distance function satisfies a set of modified axioms of metric. In the second part of the paper several feature spaces have one distance function. Galina V. Iofina was born in 1984 and graduated from the Faculty of Control and Applied Mathematics of the Moscow Institute of Physics and Technology in 2007. Upon graduating, she became a postgraduate student at the same institute with a specialization in “Foundations of Information Science.” She has been an assistant at the Department of Mathematical Foundations of Control since September of 2007 and a tutor in discrete mathematics. Her scientific interests include image recognition, discrete mathematics, optimization, mathematical statistics, and all the related fields. She is the author of 2 papers and 3 conference communications.     

Calculation and optimization of thresholds for sets of software metrics

In this article, we present a novel algorithmic method for the calculation of thresholds for a metric set. To this aim, machine learning and data mining techniques are utilized. We define a data-driven methodology that can be used for efficiency optimization of existing metric sets, for the simplification of complex classification models, and for the calculation of thresholds for a metric set in an environment where no metric set yet exists. The methodology is independent of the metric set and therefore also independent of any language, paradigm or abstraction level. In four case studies performed on large-scale open-source software metric sets for C functions, C+ +, C# methods and Java classes are optimized and the methodology is validated.     

Ensemble of feature sets and classification algorithms for sentiment classification

In this paper, we make a comparative study of the effectiveness of ensemble technique for sentiment classification. The ensemble framework is applied to sentiment classification tasks, with the aim of efficiently integrating different feature sets and classification algorithms to synthesize a more accurate classification procedure. First, two types of feature sets are designed for sentiment classification, namely the part-of-speech based feature sets and the word-relation based feature sets. Second, three well-known text classification algorithms, namely na?¨ve Bayes, maximum entropy and support vector machines, are employed as base-classifiers for each of the feature sets. Third, three types of ensemble methods, namely the fixed combination, weighted combination and meta-classifier combination, are evaluated for three ensemble strategies. A wide range of comparative experiments are conducted on five widely-used datasets in sentiment classification. Finally, some in-depth discussion is presented and conclusions are drawn about the effectiveness of ensemble technique for sentiment classification.     

A genetic classification method for speaker recognition

Gaussian mixture model (GMM) has been widely used for modeling speakers. In speaker identification, one major problem is how to generate a set of GMMs for identification purposes based upon the training data. Due to the hill-climbing characteristic of the maximum likelihood (ML) method, any arbitrary estimate of the initial model parameters will usually lead to a sub-optimal model in practice. To resolve this problem, this paper proposes a hybrid training method based on genetic algorithm (GA). It utilizes the global searching capability of GA and combines the effectiveness of the ML method. Experimental results based on TI46 and TIMIT showed that this hybrid GA could obtain more optimized GMMs and better results than the simple GA and the traditional ML method.     

A novel SVM+NDA model for classification with an application to face recognition

Support vector machine (SVM) is a powerful classification methodology, where the support vectors fully describe the decision surface by incorporating local information. On the other hand, nonparametric discriminant analysis (NDA) is an improvement over LDA where the normality assumption is relaxed. NDA also detects the dominant normal directions to the decision plane. This paper introduces a novel SVM+NDA model which can be viewed as an extension to the SVM by incorporating some partially global information, especially, discriminatory information in the normal direction to the decision boundary. This can also be considered as an extension to the NDA where the support vectors improve the choice of k-nearest neighbors on the decision boundary by incorporating local information. Being an extension to both SVM and NDA, it can deal with heteroscedastic and non-normal data. It also avoids the small sample size problem. Moreover, it can be reduced to the classical SVM model, so that existing softwares can be used. A kernel extension of the model, called KSVM+KNDA is also proposed to deal with nonlinear problems. We have carried an extensive comparison of the SVM+NDA to the LDA, SVM, heteroscedastic LDA (HLDA), NDA and the combined SVM and LDA on artificial, real and face recognition data sets. Results for KSVM+KNDA have also been presented. These comparisons demonstrate the advantages and superiority of our proposed model.     

A study of several metrics for programming effort

As the cost of programming becomes a major component of the cost of computer systems, it becomes imperative that program development and maintenance be better managed. One measurement a manager could use is programming complexity. Such a measure can be very useful if the manager is confident that the higher the complexity measure is for a programming project, the more effort it takes to complete the project and perhaps to maintain it. Until recently most measures of complexity were based only on intuition and experience. In the past 3 years two objective metrics have been introduced, McCabe's cyclomatic number v(G) and Halstead's effort measure E. This paper reports an empirical study designed to compare these two metrics with a classic size measure, lines of code. A fourth metric based on a model of programming is introduced and shown to be better than the previously known metrics for some experimental data.     

粗糙集与模糊系统集成的化学模式分类方法及其应用

模糊方法是一种有效的化学模式分类方法,但模糊规则的获取和相关参数的确定较为困难。对此,本文采用粗糙集方法,无需任何先验知识,约简系统,获取最简规则集,在此基础上构建结构合理．适用于分类的模糊-神经网络系统,并根据规则的统计性质和离散化结果初始化网络参数,采用LM方法训练网络;在橄榄油模式分类建模的应用中,该方法训练收敛速度快,所建模型预测性能良好,要优于现代统计方法和前馈神经网络。     

Feature sub-set selection metrics for Arabic text classification

Feature sub-set selection (FSS) is an important step for effective text classification (TC) systems. This paper presents an empirical comparison of seventeen traditional FSS metrics for TC tasks. The TC is restricted to support vector machine (SVM) classifier and only for Arabic articles. Evaluation used a corpus that consists of 7842 documents independently classified into ten categories. The experimental results are presented in terms of macro-averaging precision, macro-averaging recall and macro-averaging F₁ measures. Results reveal that Chi-square and Fallout FSS metrics work best for Arabic TC tasks.     

应用软件质量工程中过程度量的研究

过程度量是软件开发过程中实施软件质量保证(SQA)的一个重要课题。文章对过程度量的概念进行了一些介绍,讨论了在开发应用软件过程中常用的几种度量,可供软件开发部门在实践SQA时参考。     

Increasing the efficiency of combinatorial logical data analysis in recognition and classification problems

Problems of increasing the efficiency of combinatorial logical data analysis in recognition problems are examined. A technique for correct conversion of initial information for reduction of its dimensionality is proposed. Results of testing this technique for problems of real medical prognoses are given. Djukova Elena V. Born 1945. Graduated from Moscow State University in 1967. Candidate’s degree in Physics and Mathematics in 1979. Doctoral degree in Physics and Mathematics in 1997. Dorodnicyn Computing Center, Russian Academy of Sciences, leading researcher. Moscow State University, lecturer. Moscow Pedagogical University, lecturer. Scientific interests: discrete mathematics and mathematical method of pattern recognition. Author of 70 papers. Peskov Nikolai V. Born 1978. Graduated from Moscow State University in 2000. Candidate’s degree in 2004. Dorodnicyn Computing Center, Russian Academy of Sciences, junior researcher. Scientific interests: discrete mathematics and mathematical methods of pattern recognition. Author of ten papers. Inyakin Andrey S. Born 1978. Graduated from Moscow State University in 2000. Dorodnicyn Computing Center, Russian Academy of Sciences, junior researcher. Scientific interests: discrete mathematics and mathematical methods of pattern recognition. Author of ten papers. Sakharov Aleksei A. Born 1980. Graduated from Moscow State University in 2003. Moscow Pedagogical University, graduate student. Scientific interests: discrete mathematics and mathematical method of pattern recognition. Author of three papers.     

A general feature-weighting function for classification problems

Feature weighting is a vital step in machine learning tasks that is used to approximate the optimal degree of influence of individual features. Because the salience of a feature can be changed by different queries, the majority of existing methods are not sensitive enough to describe the effectiveness of features. We suggest dynamic weights, which are dynamically sensitive to the effectiveness of features. In order to achieve this, we propose a differentiable feature weighting function that dynamically assigns proper weights for each feature, based on the distinct feature values of the query and the instance. The proposed weighting function, which is an extension of our previous work, is suitable for both single-modal and multi-modal weighting problems, and, hence, is referred to as a General Weighting Function. The number of parameters of the proposed weighting function is fewer compared to the ordinary weighting methods. To show the performance of the General Weighting Function, we proposed a classification algorithm based on the notion of dynamic weights, which is optimized for one nearest neighbor algorithm. The experimental results show that the proposed method outperforms the ordinary feature weighting methods.     

Analysis of finite mixture of distributions: a statistical tool for biological classification problems

A microcomputer program to analyze finite mixtures of normal, binomial, Poisson or exponential distributions by a maximum likelihood estimation procedure is described. The program is coded in Turbo Pascal. Some theoretical and practical aspects of the compound distributions are discussed, mainly mathematical characteristics, fitting procedures and tests of hypotheses. The statistical tool, which is a cluster analysis technique, is presented in a general context for applications in biology. In particular an ecological example is briefly described: the age class structure resolution of a white-tailed deer population. To improve the usefulness for classification purposes, the link with discriminant analysis is examined. For a successful analysis of finite mixture of distributions the need for a large sample size is emphasized.