Hindawi Publishing CorporationEURASIP Journal on Bioinformatics and Systems BiologyVolume 2007, Article ID 79879, 9 pagesdoi:10.1155/2007/79879Research ArticleInformation-Theoretic Inference of Large TranscriptionalRegulatory NetworksPatrick E. Meyer, Kevin Kontos, Frederic Lafitte, and[r]
main reasons to keep the dimensionality of the input featuresas small as possible: computational cost and classificationaccuracy. It has been observed that added irrelevant featuresmay actually degrade the performance of classifiers if thenumber of training samples is small relative to the numb[r]
Hindawi Publishing CorporationEURASIP Journal on Applied Signal ProcessingVolume 2006, Article ID 30274, Pages 1–11DOI 10.1155/ASP/2006/30274Information Theory for Gabor Feature Selection forFace RecognitionLinlin Shen and Li BaiSchool of Computer Sc ience and[r]
and Eric Hoffman31Department of Electrical and Computer Engineering, Virginia Polytechnic Institute and State University, Arlington, VA 22203, USA2Department of Pediatric Oncology, National Cancer Institute, Gaithersburg, MD 20877, USA3Research Center for Genetic Medicine[r]
Using the formula for the differential entropy of a transformation as given in (5.13)of Chapter 5, we obtain a corresponding result for mutual information. We havefor an invertible linear transformation:(10.5)Now, let us consider what happens if we constrain the to be unc[r]
where the constant term does not depend onB. This shows the fundamental relationbetween negentropy and mutual information.We see in (10.7) that finding an invertible linear transformationBthat minimizesthe mutual information is roughly equivalent to finding direction[r]
guage applications such as machine translation, in-formation retrieval, and speech processing (Ide andV´eronis, 1998). Almost all of sense disambigua-tion methods are heavily dependant on manuallycompiled lexical resources. However these lexicalresources often miss domain specific word senses,[r]
EURASIP Journal on Applied Signal Processing 2003:10, 1001–1015c 2003 Hindawi Publishing CorporationWatermarking-Based Digital AudioData AuthenticationMartin SteinebachFraunhofer Institute IPSI, MERIT, C4M Competence for Media Security, D-64293 Darmstadt, GermanyEmail: martin.steinebach@ipsi[r]
this we proposed two techniques for featurereduction based on word clustering and se-lection. A number of word similarity mea-sures are proposed for clustering words forthe Named Entity Recognition task. A fewcorpus based statistical measures are used forimportant word select[r]
ing how evidence in the primary linguistic datacan account for first language acquisition by in-fant children (Finch and Chater, 1992a; Finch andChater, 1992b; Redington et al., 1998). At thisearly phase of learning, only limited sources ofinformation can be used: primarily distributio[r]
using concept vector space model.A document is regarded as a conglomerate con-cept that comprises by many concepts. Hence, an n-dimensional concept vector space model is defined insuch a way that a document is recognized as a vec-tor in n-dimensional concept space. We used lexicalchains for th[r]
models for sound source amplitudes and vehicles during theBayesian filtering. We used a crude form of DOA likelihoodfunction. More accurate, experimentally validated character-ization could also help improve the tracking.Reliability is an important issue in the sensor network.A single l[r]
encounter the problems, mentioned above, which arise with MT output. 7 Conclusion We present an integrated approach to extract the named entities from machine translated text, using name entity information from both source and target language. Our experiments show that with a combinati[r]
expressed as a proportion of all possible linkages — are some of the mainparameters of network information that are useful for management of groups.A network showing large structural holes indicates either a less cohesive groupor a diverse group in terms of functional or cultural backg[r]
we actually developed these resuks in basically the reverse order. Concordance analysis is stilt extremely labor-intensive, and prone to errors of omission. The ways that concordances are sorted don't adequately support current lexicographic practice. Despite the fact that a concordance is in[r]
What the EMH Does and Does NOTSayInvestors can throw darts to select stocks.This is almost, but not quite, true.An investor must still decide how risky a portfolio hewants based on risk aversion and the level ofexpected return.Prices are random or uncaused.Prices reflect info[r]
categories are related to the subject content, e.g., sport or education. Yang and Pedersen (1997) investigate five FS metrics and report that good FS methods improve the categorization accuracy with an aggressive feature removal using DF, IG, and CHI. More recently, Forma[r]
cated on each side bet ween longitudinal yellow lines. When mature, the nearly 2-inch caterpillars move down the tree trunks to spin their white, silken cocoons on the bark of trees, on buildings, in grass and in other sheltered loca-tions. The caterpillars can be quite noticeable as they cro[r]
health care and biomedicine New York, NY: Springer, 3 2006.2. Greenes RA, Shortliffe EH: Commentary: Informatics in biomedicine andhealth care. Acad Med 2009, 84:818-820.3. Bernstam EV, Hersh WR, Johnson SB, Chute CG, Nguyen H, Sim I, Nahm M,Weiner MG, Miller P, DiLaura RP, Overcash M, Lehman[r]