Automated feature selection is important for text categorization to reduce feature size and to speed up learning process of classifiers. In this paper, we present a novel and efficient feature ...selection framework based on the Information Theory, which aims to rank the features with their discriminative capacity for classification. We first revisit two information measures: Kullback-Leibler divergence and Jeffreys divergence for binary hypothesis testing, and analyze their asymptotic properties relating to type I and type II errors of a Bayesian classifier. We then introduce a new divergence measure, called Jeffreys-Multi-Hypothesis (JMH) divergence, to measure multi-distribution divergence for multi-class classification. Based on the JMH-divergence, we develop two efficient feature selection methods, termed maximum discrimination (<inline-formula><tex-math notation="LaTeX">MD</tex-math> <inline-graphic xlink:type="simple" xlink:href="he-ieq1-2563436.gif"/> </inline-formula>) and methods, for text categorization. The promising results of extensive experiments demonstrate the effectiveness of the proposed approaches.
Intuitive Probability and Random Processes using MATLABr is an introduction to probability and random processes that merges theory with practice. Based on the author's belief that only "hands-on" ...experience with the material can promote intuitive understanding, the approach is to motivate the need for theory using MATLAB examples, followed by theory and analysis, and finally descriptions of "real-world" examples to acquaint the reader with a wide variety of applications. The latter is intended to answer the usual question "Why do we have to study this?" Other salient features are:
*heavy reliance on computer simulation for illustration and student exercises
*the incorporation of MATLAB programs and code segments
*discussion of discrete random variables followed by continuous random variables to minimize confusion
*summary sections at the beginning of each chapter
*in-line equation explanations
*warnings on common errors and pitfalls
*over 750 problems designed to help the reader assimilate and extend the concepts
Intuitive Probability and Random Processes using MATLABr is intended for undergraduate and first-year graduate students in engineering. The practicing engineer as well as others having the appropriate mathematical background will also benefit from this book.
About the Author
Steven M. Kay is a Professor of Electrical Engineering at the University of Rhode Island and a leading expert in signal processing. He has received the Education Award "for outstanding contributions in education and in writing scholarly books and texts..." from the IEEE Signal Processing society and has been listed as among the 250 most cited researchers in the world in engineering.
This paper develops the mathematical framework to analyze the stochastic resonance (SR) effect in binary hypothesis testing problems. The mechanism for SR noise enhanced signal detection is explored. ...The detection performance of a noise modified detector is derived in terms of the probability of detection P D and the probability of false alarm P FA. Furthermore, sufficient conditions are established to determine the improvability of a fixed detector using SR. The form of the optimal noise pdf is determined and the optimal stochastic resonance noise pdf which renders the maximum P D without increasing P FA is derived. Finally, an illustrative example is presented where performance comparisons are made between detectors where the optimal stochastic resonance noise, as well as Gaussian, uniform, and optimal symmetric noises are applied to enhance detection performance.
Three common techniques to discriminate between alternatives in a binary hypothesis testing problem are: the generalized likelihood ratio test (GLRT), the Rao test, and the Wald test. In this paper, ...we investigate some characteristics of the corresponding decision statistics and provide their expressions for some problems of particular interest in statistical signal processing. First of all, we focus on the invariance of the Rao and Wald tests with respect to transformations leaving the testing problem unaltered. Then, we introduce necessary and sufficient conditions in order for their decision statistics to coincide with twice the logarithm of the GLRT statistic. Finally, we present some detection problems, usually encountered in practical signal processing applications, where the decision variables of the three quoted tests are equivalent, namely related by strictly monotonic transformations.
A novel three-component reaction of pyridine N-oxides, acyl chlorides, and cyclic ethers is described. Treatment of an electron-deficient pyridine N-oxide with an acyl chloride in the presence of a ...cyclic ether at 25–50 °C leads to a substituted pyridine as a single regioisomer in up to 58% isolated yield. Isotopic-labeling experiments and substrate scope support the reaction proceeding through a carbene intermediate.
A new approach to the problem of dimensionality reduction is proposed. The specific application is to the detection of signals in noise, although it should be applicable to other signal processing ...problems of current interest. Using a minimum mean square error estimator of the likelihood ratio one can determine a low dimensional statistic, not necessarily linear in the data, that performs well for detection, i.e., with minimal loss of information. If a sufficient statistic does exist for the problem then the proposed approach yields the well known result that one should use the likelihood ratio of the sufficient statistic for detection. Other interesting relationships are explored and some specific examples are given.
A major challenge for kidney transplantation is balancing the need for immunosuppression to prevent rejection, while minimizing drug‐induced toxicities.
We used DNA microarrays (HG‐U95Av2 GeneChips, ...Affymetrix) to determine gene expression profiles for kidney biopsies and peripheral blood lymphocytes (PBLs) in transplant patients including normal donor kidneys, well‐functioning transplants without rejection, kidneys undergoing acute rejection, and transplants with renal dysfunction without rejection. We developed a data analysis schema based on expression signal determination, class comparison and prediction, hierarchical clustering, statistical power analysis and real‐time quantitative PCR validation. We identified distinct gene expression signatures for both biopsies and PBLs that correlated significantly with each of the different classes of transplant patients. This is the most complete report to date using commercial arrays to identify unique expression signatures in transplant biopsies distinguishing acute rejection, acute dysfunction without rejection and well‐functioning transplants with no rejection history. We demonstrate for the first time the successful application of high density DNA chip analysis of PBL as a diagnostic tool for transplantation. The significance of these results, if validated in a multicenter prospective trial, would be the establishment of a metric based on gene expression signatures for monitoring the immune status and immunosuppression of transplanted patients.
Treatment of electron deficient pyridine N-oxides with 4-nitrobenzoyl chloride and a cyclic thioether in the presence of triethylamine leads to the corresponding 2-functionalized product in up to a ...74% isolated yield. The transformation can also be accomplished with alternative nitrogen containing heterocycles, including quinolines, pyrimidines, and pyrazines. To expand the scope of the transformation, diisopropyl ether can be used as the reaction medium to allow for the use of solid thioether substrates.
Recent advances in design of powered artificial legs have led to increased potential to allow lower limb amputees to actively recover from stumbles. To achieve this goal, promptly and accurately ...identifying stumbles is essential. This study aimed to 1) select potential stumble detection data sources that react reliably and quickly to stumbles and can be measured from a prosthesis, and 2) investigate two different approaches based on selected data sources to detect stumbles and classify stumble types in patients with transfemoral (TF) amputations during ambulation. In the experiments, the normal gait of TF amputees was perturbed by a controllable treadmill or when they walked on an obstacle course. The results showed that the acceleration of prosthetic foot can accurately detect the tested stumbling events 140-240 ms before the critical timing of falling and precisely classify the stumble type. However, the detector based on foot acceleration produced high false alarm rates, which challenged its real application. Combining electromyographic (EMG) signals recorded from the residual limb with the foot acceleration significantly reduced the false alarm rate but sacrificed the detection response time. The results of this study may lead to design of a stumble detection system for instrumented, powered artificial legs; however, continued engineering efforts are required to improve the detection performance and resolve the challenges that remain for implementing the stumble detector on prosthetic legs.
It is shown in this letter that the addition of noise to data, resulting in noise enhanced detection, is equivalent to using a randomized decision rule. Since the theory of randomized decision-making ...is well developed, this link can be exploited to better understand the advantages, disadvantages, and general properties of noise enhanced detection. As an example, we show how some recent results can be interpreted within this more general framework.