Non-parametric Classification Technique Research Articles

Non-parametric distance-based classification techniques and their applications

Pattern Recognition is a key scientific field of longstanding tradition, with origins in the early years of computer science. Today, Pattern Recognition has reached a level of maturity that allows us to build highly sophisticated systems, which perform very different tasks. Nevertheless, its evolution has opened up a number of new problems, ranging from specific algorithms to system integration, which remain elusive and assure a long life for this research field. The field is progressing rapidly, and an air of excitement among researchers is being created by the increasing scope of applications to which Pattern Recognition is relevant, and by the many technical advances that have been made in recent years. One reason Pattern Recognition is such a rapidly developing field lies in the fact that modern societies have entered the ‘‘data era.’’ An unprecedented investment is being made in the collection of data, with archives being formed on an enormous scale. Biological data are being acquired using increasingly fast machines to scan genomes, hyperspectral satellite imagery is being stored on a massive scale, and web documents are appearing at an explosive rate in internet, and so on. The development of effective ways for extracting useful information from these data stores is an overall challenge to computer science as a discipline. This goal drives much of present pattern recognition and machine learning research. This Special Issue aims to provide a platform for Pattern Recognition researchers to present their newly developed techniques and applications in the area of non-parametric Pattern Recognition. From a practical point of view, nonparametric classification is one of the most relevant classification approaches for real world problems. From a theoretical point of view, it is still a very active area of research in spite of the fact of being a classical topic of the field. Areas of interest for this special journal issue included, but were not limited to the following topics: distance based classification in metric and non metric spaces, feature selection and extraction, prototype selection, editing and condensing, non supervised distance-based classification, and applications. We received 45 manuscripts. Fifteen manuscripts were finally selected for this Special Issue after being reviewed by at least three external reviewers. The accepted papers can be broadly thought of as being representative of the breadth of concerns in areas like discriminant analysis, advanced nearest neighbour classification, online learning, dissimilarity-based classification, and covering a wide range of application domains. The paper ‘‘Distance-based Discriminant Analysis Method and its Applications’’ by Kosinov and Pun presents a new approach for finding discriminative linear transformations from inter-observation distances. The proposed method is suitable for both binary and multiple category problems and allows for a kernel-based extension that overcomes the linearity assumption of the initial formulation. F. Pla (&) Department de Llenguatges i Sistemes Informatics, Universitat Jaume I, Castello, Spain e-mail: Filiberto.Pla@lsi.uji.es

Spatio-Temporal Feature-Extraction Techniques for Isolated Gesture Recognition in Arabic Sign Language

This paper presents various spatio-temporal feature-extraction techniques with applications to online and offline recognitions of isolated Arabic Sign Language gestures. The temporal features of a video-based gesture are extracted through forward, backward, and bidirectional predictions. The prediction errors are thresholded and accumulated into one image that represents the motion of the sequence. The motion representation is then followed by spatial-domain feature extractions. As such, the temporal dependencies are eliminated and the whole video sequence is represented by a few coefficients. The linear separability of the extracted features is assessed, and its suitability for both parametric and nonparametric classification techniques is elaborated upon. The proposed feature-extraction scheme was complemented by simple classification techniques, namely, K nearest neighbor (KNN) and Bayesian, i.e., likelihood ratio, classifiers. Experimental results showed classification performance ranging from 97% to 100% recognition rates. To validate our proposed technique, we have conducted a series of experiments using the classical way of classifying data with temporal dependencies, namely, hidden Markov models (HMMs). Experimental results revealed that the proposed feature-extraction scheme combined with simple KNN or Bayesian classification yields comparable results to the classical HMM-based scheme. Moreover, since the proposed scheme compresses the motion information of an image sequence into a single image, it allows for using simple classification techniques where the temporal dimension is eliminated. This is actually advantageous for both computational and storage requirements of the classifier.

Distributions of Galaxy Spectral Types in the Sloan Digital Sky Survey

We perform an objective classification of 170,000 galaxy spectra in the Sloan Digital Sky Survey (SDSS) using the Karhunen-Loeve (KL) transform. With about one-sixth of the total set of galaxy spectra that will be obtained by the survey, we are able to carry out the most extensive analysis of its kind to date. The formalism proposed by Connolly and Szalay is adopted to correct for gappy regions in the spectra and to derive eigenspectra and eigencoefficients. From this analysis, we show that this gap-correction formalism leads to a converging set of eigenspectra and KL-repaired spectra. Furthermore, KL eigenspectra of galaxies are found to be convergent not only as a function of iteration, but also as a function of the number of randomly selected galaxy spectra used in the analysis. From these data a set of 10 eigenspectra of galaxy spectra are constructed, with rest-wavelength coverage 3450–8350 A. The eigencoefficients describing these galaxies naturally place the spectra into several classes defined by the plane formed by the first three eigencoefficients of each spectrum. Spectral types corresponding to different Hubble types and galaxies with extreme emission lines are identified for the 170,000 spectra and are shown to be complementary to existing spectral classifications. From a nonparametric classification technique, we find that the population of galaxies can be divided into three classes that correspond to early late-type through intermediate late-type galaxies. This finding is believed to be related to the color separation of SDSS galaxies discussed in earlier works. Bias in the spectral classifications due to the aperture spectroscopy in the SDSS is small and within the signal-to-noise limit for a majority of galaxies, except for the reddest nearby galaxies and large galaxies (>30 kpc) with prominent emissions. The mean spectra and eigenspectra derived from this work can be downloaded from the SDSS Web site.

Handwriting style classification

This paper describes an independent handwriting style classifier that has been designed to select the best recognizer for a given style of writing. For this purpose a definition of handwriting legibility has been defined and a method implemented that can predict this legibility. The technique consists of two phases. In the feature-extraction phase, a set of 36 features is extracted from the image contour. In the classification phase, two nonparametric classification techniques are applied to the extracted features in order to compare their effectiveness in classifying words into legible, illegible, and middle classes. In the first method, a multiple discriminant analysis (MDA) is used to transform the space of extracted features (36 dimensions) into an optimal discriminant space for a nearest mean based classifier. In the second method, a probabilistic neural network (PNN) based on the Bayes strategy and nonparametric estimation of probability density function is used. The experimental results show that the PNN method gives superior classification results when compared with the MDA method. For the legible, illegible, and middle handwriting the method provides 86.5% (legible/illegible), 65.5% (legible/middle), and 90.5% (middle/illegible) correct classification for two classes. For the three-class legibility classification the rate of correct classification is 67.33% using a PNN classifier.

Molecular abnormalities in the major psychiatric illnesses: Classification and Regression Tree (CRT) analysis of post-mortem prefrontal markers.

Post-mortem specimens from the Stanley Foundation Neuropathology Consortium, which contains matched samples from patients with schizophrenia, bipolar disorder, non-psychotic depression and normal controls (n = 15 per group), have been distributed to many research groups around the world. This paper provides a summary of abnormal markers found in prefrontal cortical areas from this collection between 1997 and 2001. With parametric analyses of variance of 102 separate data sets, 14 markers were abnormal in at least one disease. The markers pertained to a variety of neural systems and processes including neuronal plasticity, neurotransmission, signal transduction, inhibitory interneuron function and glial cells. The data sets were also examined using the non-parametric Classification and Regression Tree (CRT) technique for the four diagnostic groups and in pair-wise combinations. In contrast to the results obtained with analyses of variance, the CRT method identified a smaller set of nine markers that contributed maximally to the diagnostic classifications. Three of the nine markers observed with CRT overlapped with the ANOVA results. Six of the nine markers observed with the CRT technique pertained to aspects of glutamatergic, GABA-ergic, and dopaminergic neurotransmission.

Classification via kernel product estimators

Multivariate kernel density estimation is often used as the basis for a nonparametric classification technique. However, the multivariate kernel classifier suffers from the curse of dimensionality, requiring inordinately large sample sizes to achieve a reasonable degree of accuracy in high dimensional settings. A variance stabilising approach to kernel classification can be motivated through an alternative interpretation of linear and quadratic discriminant analysis in which rotations of the coordinate axes are employed to obtain an assumed mutual independence among the components of the rotated data. This alternative method, which we call the method of kernel product estimators, performs well in a variety of examples, including a 20-dimensional target recognition problem.

Goal-directed classification using linear machine decision trees

Recent work in feature-based classification has focused on nonparametric techniques that can classify instances even when the underlying feature distributions are unknown. The inference algorithms for training these techniques, however, are designed to maximize the accuracy of the classifier, with all errors weighted equally. In many applications, certain errors are far more costly than others, and the need arises for nonparametric classification techniques that can be trained to optimize task-specific cost functions. This correspondence reviews the linear machine decision tree (LMDT) algorithm for inducing multivariate decision trees, and shows how LMDT can be altered to induce decision trees that minimize arbitrary misclassification cost functions (MCF's). Demonstrations of pixel classification in outdoor scenes show how MCF's can optimize the performance of embedded classifiers within the context of larger image understanding systems.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

A Method of Fault Signature Extraction for Improved Diagnosis

Efficient extraction of fault signatures from sensory data is a major concern in fault diagnosis. This paper introduces a self-tuning method of fault signature extraction that enhances fault detection, minimizes false alarms, improves diagnosability, and reduces fault signature variability. The proposed method uses a Flagging Unit to convert the processed measurements to binary vectors, and utilizes nonparametric pattern classification techniques to estimate the fault signatures. The performance of the Flagging Unit, which relies on its adaptation algorithms to optimize its performance based upon a sample batch of measurement-fault vectors, is demonstrated in simulation.

Comparison of two classification approaches for automatic density separation of Florida citrus

A non-parametric classification technique was utilized to develop breakpoints for density separation of freeze-damaged citrus and compared to Gaussian distribution results. Class probability and economic loss factors were included in both models. The techniques were applied to density sorting based on optical dimensional sizing coupled with real-time weight measurements. A major change in density occurs in citrus when exposed to freezing conditions after which internal desic-cation can be correlated to a reduction in the density of the fruit. For a three-category separation, density breakpoints ranged from 0.80 to 0.90, non-parametric, and 0.80 to 0.98, Gaussian.

A statistical analysis of prescreening alarms in a population of normal and abnormal gynecologic specimens.

A multidimensional slit-scan flow system has been developed to serve as an automated prescreening instrument for gynecological cytology. Specimens are classified abnormal based on the number of cells having elevated nuclear fluorescence (alarms). An alarm region in a bivariate histogram of nuclear fluorescence versus nuclear-to-cell-diameter ratio is defined. Alarm region probability arrays are calculated to estimate the probability that an alarm falling in a particular bin of the alarm region is either from a normal or an abnormal specimen. From these arrays, a weighted alarm index is generated. In addition, summary indices are derived that measure how the distribution of alarms in each specimen compares with the average distributions for the normal and abnormal specimen populations. These indices together with current features are evaluated with respect to their utility in specimen classification using a nonparametric classification technique known as recursive partitioning. Resulting classification trees are presented that suggest information in the distribution of alarms in the bivariate histogram. In addition, they validate the features and rules currently used for specimen classification. Recursive partitioning appears to be useful for multivariate classification and is seen as a promising technique for other applications.

An overview of pattern recognition and its application in acoustics

The problem of pattern recognition is introduced and its major approaches briefly reviewed. In the decision‐theoretic approach, both parametric and nonparametric classification techniques are described. In the syntactic approach, pattern description languages and their recognition algorithms are discussed. Application examples are presented. VLSI implementations of some of the techniques are discussed. A unified syntactic‐semantic approach is then proposed. Problems for further study are suggested. [Work supported by NSF under Grant ECS 81‐19886.]

Non-Parametric Methods for Detection of Adulteration of Concentrated Orange Juice for Manufacturing

Abstract The composition of organic constituents (total sugars, reactive phenols, total amino acids, arginine, and γ-aminobutyric acid) has been measured in a large (360 samples) selection of concentrated orange juice for manufacturing and orange pulp wash in the U.S. trade. The detection of adulteration with sugar, reducing sugars, and citric acid addition has been investigated by using non-parametric nearest neighbor classification techniques in the 4-space of log ratios of the compositions. The results show that such detection is possible with a type 1 = type 2 error rate of 1 0% for 2 0% adulteration if at least 7 samples are taken. The assumptions of such samplings are discussed.