Morphological Ending based Strategies of Unknown Word Estimation for Statistical POS Urdu Tagger
|
|
- Angelina Potter
- 5 years ago
- Views:
Transcription
1 IJCSES Iteratoal Joural of Computer Sceces ad Egeerg Systems, Vol., No.3, July 2007 CSES Iteratoal c2007 ISSN Morphologcal Edg based Strateges of Ukow Word Estmato for Statstcal POS Urdu Tagger Waqas ANWAR, Xua WANG, Lu LI 2, ad Xao-Log WANG 2 School of Computer Scece ad Techology,Harb Isttute of Techology Shezhe Graduate School Shezhe, 58055, R.P.Cha E-mal:{waqas,xuawag}@su.ht.edu.c 2 School of Computer Scece ad Techology,Harb Isttute of Techology Shezhe Graduate School Shezhe, 58055, R.P.Cha E-mal:{ll,wagxl}@su.ht.edu.c Abstract Natural laguage processg has wdely used Statstcal based laguage models to solve dsambguato problems. Over the past decades dfferet techques regardg POS taggg have bee proposed for Eglsh, Europea ad East Asa laguages. I ths paper our focus s POS taggg for Urdu due to the facy stage of Urdu laguage based taggg system. We have combed two approaches (Statstcal ad morphologcal edg based techque) to assg the approprate sytactc categores, usg Urdu laguage corpus as our expermetal data. The process of our taggg system cossts of two dfferet stages. I frst stage, we apply statstcal based word level laguage model to compute the probablty ad assg the approprate tag to the word. I secod stage, we extract all the ambguous ad ukow words the corpus ad apply morphologcal edg based rules to resolve these dscotutes, whch arse f statstcal model fals to assg the approprate tag to a gve word. The developmet of ths tagger s a tal step toward the Urdu POS taggg. The expermetal results of the tagger show that the performace of the ukow word s mproved whe we add morphologcal edg based features wth statstcal model. Evaluato method employed shows the sgfcace of expermetal results ad the effectveess of morphologcal edg o statstcal method. Ths s a poeerg work towards buldg a POS tagger for Urdu laguage through Morphologcal edg based strateges Key words: Urdu Laguage, Morphologcal Based Edg Model, Probablstc Model, Part-of-speech Taggg.. Itroducto Research automatc text taggg s a fudametal ad mportat task atural laguage processg, whch s cosdered to be solved for most of the corpus rch Laguages. O the other had uder-prvlege resources laguage the problem of automatc text taggg s stll ts tal stage especally Urdu laguage. The POS taggg provdes prelmary step towards buldg varous atural laguage processg applcatos e.g. Iformato extracto, parsg, word sese dsambguato, text-to-speech coverso etc I Eglsh ad other smlar laguages may mportat results have bee acheved by applyg statstcal based methods. However for the Laguages whch exhbt dfferet behavor tha Eglsh lke Urdu, Hd, Turksh ad Czech, may of the successful techques appear capactated whle solvg the taggg problem [9],[0]. Several dfferet approaches have bee used for automatc part of speech taggg. Geerally t ca be classfed to two dfferet groups: rule based approach ad statstcal approach. But most of the free word ad hghly flected laguages the researcher solve dsambguato problem through hybrd approach. Because morphologcal edg based approach play a vtal role to solve ukow word problem. Eve at some pot the researcher use frst edg based techque ad the use statstcal based techque. Here we wll dscuss some of the tal work about features based techque as secodary strategy wth combato of statstcal method. Itally (Churcg, 989) [2] used the captalzato features to recogze the ukow proper ous that occur durg trag process. I (Meteer et al, 99) pot out that the accuracy rate ad specfcally ukow words accuracy sgfcatly mprove after addg morphologcal edg based Kowledge. They defe dfferet features such as flecto, dervatoal edgs, hypheato ad captalzato. The word edg characterstcs vares from laguage to laguage ad requred laguage kowledge. They also report that the morphologcal edg of the words reduces the error rate o the ukow word by a factor of 3[4]. I (Adams ad Nefeld, 993) arbtrary character edgs s appled alog wth some other mportat laguage features such as captalzato ad puctuato [2]. Also there taggg system uses exteral lexco that s prepared wthout statstcal formato but has POS tag wth every word to assg the correct tag to the word. I (Adam el al, 994) motvates the dea that Mauscrpt receved May, Mauscrpt revsed Jue 5, 2007.
2 68 IJCSES Iteratoal Joural of Computer Sceces ad Egeerg Systems, Vol. No.3, July 2007 prmarly use the edg based strategy ad secodary use whole-word statstcs based approach [],[2]. Oe feels that laguages lke Urdu has ot bee studed -depth to corporate statstcal methods as compared to other laguages. Urdu s the wdely spoke laguage South Asa. It s morphologcally rch ad free word laguage. Though, the part of speech taggg Urdu laguage s stll ts facy stage. The reasos for workg o part of speech taggg Urdu laguage are the followg: The amout of avalable aotated lgustcs resources for Urdu laguage s much smaller tha for well researched laguages. No well-defed Urdu part of speech tag set s avalable that cotas all the formato. Comprehesve Urdu part of speech lexco s very dffcult to fd. All the above meto reasos are major hdrace Urdu laguage processg research ad accout for the fact, why Urdu laguage processg s lackg behd [6]. I fact Urdu s hghly flected laguage ad the lmtatos of the statstcal based methods may be recompesed wth the stregth of rule based methods. Aother problem wth part of speech taggg of Urdu laguage s how to utlze ad alter exstg techques to reap maxmum effcecy ad to detfy the best method to be used cojucto wth lmted aotated corpora. Keepg ths vew we have developed a edg strateges based o statstcal model to fully utlze the correspodg features of Urdu laguage. The ma cotrbuto of ths work s to detfy the techques ad laguage models that are approprate for desgg a Urdu POS taggg system. To the best of our kowledge we are the frst to deal wth POS taggg task for Urdu laguage through ths techque. I ths paper, we wll descrbe morphologcal edg based part-of-speech taggg model for Urdu laguage. The frst step of our model s to compute word level statstcs. The secod step of our model s to add the morphologcal edg-based features. However, our statstcal model s based o three methods (ugram, bgram, backoff). Durg the mplemetato of our statstcal model we observed that the basele algorthm of our tagger does ot performed well. Due to the lmtato statstcal mode ad laguage characterstcs we combe our statstcal based model wth morphologcal edg based features. The Expermetal results of our tagger show that the better performace ca be acheved by combg statstcal model wth morphologcal features stead of pure statstcal model. By applyg these morphologcal features the overall performace crease up to.2% but ukow word error rate s sgfcatly decreased. Ths paper s orgazed as follow: Secto 2 gves a bref overvew of Urdu laguage, Secto 3 descrbes the taggg model (ugram, bgram ad backoff).i ths secto we also descrbes the detal of our two taggg algorthms. Secto 4 presets the expermetal results. Secto 5 presets the model evaluato. Cocluso ad future work are preseted at the ed 2. A Bref Overvew of Urdu Laguage Urdu s a dervatoal word from Turksh ad ts mea horde (Lashkar). Urdu, a Ido-Europea laguage of the Ido Arya famly, s spoke Ida ad Paksta. It s the atoal laguage of Paksta havg eleve mllo speakers. Amog all the laguages the world t s most closely smlar to Hd laguage. Urdu ad Hd both have orgated from the dalect of Delh rego ad other tha the mute detals these laguages share ther morphology. Sce Hd has adopted may words from Saskrt, Urdu has borrowed a large umber of vocabulary tems from Persa ad Arabc Urdu s also borrowg umber of vocabulary from Turksh, Portuguese ad Eglsh. May of the Arabc words have bee borrowed by Urdu laguage through Persa laguage. These words vary slghtly ther toe, cootatos ad feelg. Oe of the oteworthy aspects of Urdu grammar costtuto s ts word order SOV (subject, object, ad verb). Ths order does exhbt some flexblty as the subject proous are frequetly dropped [2]. 3. Laguage Model I atural laguage processg statstcal laguage model s employed to calculate ad assg a probablty Pw ( ) to every word ad to decde the accurate target word sequece w= w, w2,... w. Let s say that gve a partcular word sequece w we wat to fd the most probable tag sequecet. I our algorthm we compute the probablty of dvdual word. Weather a gve tag t s
3 Morphologcal Edg based Strateges of Ukow Word Estmato for Statstcal POS Urdu Tagger 69 approprate for a gve word w ca be aswered by fdg the probablty Pwt (, ). P( wt, ) = P( w) P( t w) () P( t w ) = P( t, t2,... t t0, w, w2,... w) = t0, w, w,... w). 2 t, t0, w, w2,... w)... t... t0, t, w, w2,... w)... pt ( t,..., t, t0, w, w2,... w) = t0, w). 2 t, w2)... t, w)... t, w) P( t w ) = P( t t, w) (2) = I equato () Pws ( ) the ucodtoal probablty ad detcal for all tag sequeces ad that s why we gored Pw. ( ) w) s the codtoal probablty of the occurreces of sequece t gve that a sequece of word w occurred. I ths case we ca make the assumpto that each tag deped oly the mmedate precedg tag wth curret word ad curret tag (bgram model). We called ths s a Markov assumpto. The model use ths paper s dfferet from the oe use [4]. I [4] the authors used the geeratve model to solve the dsambguato problem. The Geeratve dsambguato s a model of the jot probablty P (W,T),where W represets the words ad T represets the specfc tag set as put. I ths model we make the assumpto through Bayes rules to compute the P (T W) ad to fd out most probable tag sequece T.[3] O the other had our model s based o Dscrmatve model or codtoal probablty. I ths model we ca compute P (T W) drectly Because W s already kow as put. The other researchers also pot out covcg causes regardg the use of dscrmatve rather tha geeratve model. I fact, f we leave the ssue of ukow word hadlg whch s more mportat dsambguato problem the dscrmatve dsambguato model s mostly preferred over geeratve dsambguato model. [3] The word sequece probablty value ca be calculated from the tagged corpus whch s obta from trag corpus usg, w) w) (3) Cw ( ), t, w) t, w) (4), w) * t t = arg max P( t w ) (5) t = arg max P( t t w ) (6) * t Cw ( ) Occurreces of the word w Ctw (, ) Occurreces of tag t tagged wth w ) Occurreces of the tag t, t2) Occurreces of the tag bgram t, t 2 tw ) Occurreces of the tag-tag word bgram tw ) that s the tag t followed by the tag t word wth w I our approach ugram tagger s based o covetoal statstcal taggg algorthm whch commts a tag for each toke that s most probable for the word questo. Whle determg the probablty of the most probable tag for a toke we must take to cosderato the relatve frequeces of the tags desgated to the toke. These data set are ot mache tagged rather ths procedure requres huma terveto. Addtoally maxmum lkelhood estmato s used to tag our taggers..i case of ugram taggg the optmal dctoary lst out all words w, w2,..., w ad ther correspodg tags t, t 2,... t.the words ad tags are dcated by w adt. Thus ma task of the tagger s to fd out the best Part of speech tag * t wth maxmzg the codtoal probablty w ) [], [3] [5], [7]. Whle ugram model requred most probable tags, bgram model eeds dual formato. It collects tags for the curret word ad ts precedg tag to calculate the cotext for the toke. Ths cotext s ter used by the tagger to assg most probable tag. Maxmum lkelhood estmato s also employed to tag the data. Table shows buld up of a dctoary wth bgram model. Ths dcates prevous tag wth curret word agast curret word. Ths dctoary s optmal because t cludes all the word ad correspodg tags wth the avalable corpus. 3. Back-off model The Backoff model essetally overcomes the sparse occurreces problems the trag corpus. Due to the lack of eough amout of trag corpus word frequeces dstrbuto s msrepreseted.[8] As we already metoed our bgram laguage model s classfed by the probablty of curret tag based o curret word
4 70 IJCSES Iteratoal Joural of Computer Sceces ad Egeerg Systems, Vol. No.3, July 2007 ad prevous tag. I our algorthm case of bgram model, f the order of word par (cotextual formato) s ot foud wth the defte cotext the trag corpus, the bgram tagger s backed off to the ugram tagger. Also, f the order of the word has too sparse occurrece, the ugram tagger s backed off to the default tagger.the default tagger specfy the etre toke that are ambguous ad ukow. After that, by default assg the same tag to every toke that tag s specfy by the user []. Thus, the Bgram tagger probablty for a word w s defed as: Ctt ( w) Ctt ( w) > U ( ) Ct w) t w = Otherwse w) Procedure Bgram Back-off ( ) U: = Step: Compute Ctt ( w) Step 2: Compute w) Step 3: for to f Ctt ( w) > U, t, w) t, w) =, w) else t w) = w) Ed for Step 4: t * = arg max P( t t w) t Step 5: stop Procedure Our proposed backoff tagger gves better effcecy to the trag corpus by elmatg detcal tag smlar cotext calculated by bgram ad ugram backoff tagger durg trag process. Ths eables bgram tagger to keep mmal umber of staces. Ths umber ca also be user defed as cutoff value []. The cutoff techque of trag corpus prug elmates those ugrams ad bgrams that occur rarely wth the laguage model [8] 3.2 Morphologcal edg strateges-based model I frst stage of our tagger, the statstcal model does ot successfully tag all the words specally ukow words. Also some of the kow words are wrogly tagged. Urdu s oe of the partally free word ad hghly flected laguage. Most of these types of laguages should use orthographc rules to resolve the problem dscussed above. Lgustcally the flectoal chage drectly flueces to the word edg. I secod stage of our taggg system we apply morphologcal edg based strateges to resolve the ukow ad ambguous words those ot hadled by statstcal method. The morphologcal edg based method cossts of umber of feature set. I features set we defe all the sytactc categores rules. Such as, the ou ca be categorzed to dfferet type of character edgs. ),(و, wa o ),(يا, ya ),(وں, w ),(ی, ye ),(ع, a ),(اں, a ),(ہ, he ),(ا, alf ).[ 5 ](ں, u Smlarly we also defe all the other sytactc categores rules ad covert these rules to model form. [6].These rules also cover all the sytactcally affected flecto e.g. ou ca be flected as umber geder ad case. I fgure 3 we defe some of the most frequet edg that has occurred durg taggg process. Each rule s descrbed as character edg wth approprate tag. The character edg sze s ot fxed ad all the edg depeds o the Urdu laguage grammar characterstcs. Durg the model buldg process we extract all the ambguous ad ukow word that has occurred durg the statstcal based model taggg process. After extractg all the words we evaluate all the words edg oe by oe. I morphologcal rules mplemetato the sequece of rules play a mportat role because the sequece of rules depeds o the most frequet error occurrg tags. I our taggg system the frequet error occurrg tags are ou, adjectve, verb etc.the process of choosg the most approprate tag depeds o how well the rules are defed. But whe we mplemet these rules to our tagger the frst we estmate morphologcal edg wth o-frequet error occurrg tags ad at last we estmate the most-frequet error occurrg tags. After ths process we decde that whch tag s most approprate for each word. f Codto s true P( t codto) = 0 Otherwse 4. Expermets I the expermet, we use EMILLE corpus to test our results. The EMILLE corpus was released o 2004 by Lacaster Uversty. Ths corpus has three ma parts (moologal, aotated, parallel) [2].We evaluates the performace of our taggg system to compare two dfferet techques (statstcal ad statstcal wth combato of morphologcal edg). The result of our tagger depeds o three stadardze measures (Precso, Recall, F-measure) [4] ad these measures gve us comprehesve results about tagger performace.
5 Morphologcal Edg based Strateges of Ukow Word Estmato for Statstcal POS Urdu Tagger 7 correct umber of toke tag par occurrece Precso = total umber of toke tag par correct umber of toke tag par occurrece Recall = umber of correct toke tag par that s possble 2*Precso*Recall F-measure = Precso+Recall As we metoed above that our tagger s based o two techques. Fgure ad Fgure 2 demostrate the learg bars usg dfferet sze of trag corpus. The learg bars of the Fgure ad Fgure 2 represet the overall accuracy uder dfferet expermetal codtos (ugram, bgram, backoff) wth morphologcal ad wthout morphologcal edgs respectvely. The otable aspect both Fgures s that after applyg morphologcal edg based strateges the overall performace of the tagger creases to approxmately.2%. The expermetal results suggest that partally free word ad hghly flected laguages, the laguage features has a very mportat role. So geerally speakg morphologcal edg based strateges have better complace tha pure statstcal method. I case of the pure statstcal based model the accuracy rates of ugram ad bgram crease from 79% to 94.30% ad 6.50% to 88.50% respectvely. Here, the backoff model overcomes the sparse data problem ad the accuracy of the tagger s creased to 95.0%.Becuase trag data s ot suffcet eough so the ukow word accuracy of the pure statstcal based model s ot hgher as the case of other laguages. We overcome ths performace degradg drawback by combg the Urdu morphologcal edg based kowledge. BY usg morphologcal based kowledge cojucto, the accuracy rates of ugram ad bgram crease from 90% to 95.0% ad 8% to 93% respectvely. Also whe we apply backoff model through combato of morphologcal edg based strateges the accuracy creases from 90.60% to 96.3%. The ma cotrbutor overall accuracy mprovemet through morphologcal edg based strateges s the laguage features set. I Table 3 we represet the comprehesve result comparso betwee both of the models. I ths table we represet the Overall Precso, Overall Recall, F-Measure, ad kow overall Precso. Ukow word hadlg wth dfferet laguage features s aother mportat characterstc of our tagger. Fgure 3 shows the error rate of ukow word. I ths Fgure we cosder two aspects: oe s sze of trag corpus ad the other s laguage features. We apply the laguage feature set step by step. If we compare pure statstcal method ad statstcal method wth oly ou features there s a bg error rate dfferece. After addg the ou features we troduce more features whle creasg the sze of trag corpus at the same tme. Fgure 3 also shows that after addg all features the ukow word error rate s sgfcatly reduced.
6 72 IJCSES Iteratoal Joural of Computer Sceces ad Egeerg Systems, Vol. No.3, July 2007 R 0 / R [ ( )] *00 r0 = r Where R 0 s error dex for gve model relatve to the base model (statstcal). The Model wth ou features despte 33.44% decrease error relatve to base model. Smlarly model wth (Nou + Adjectve) features ad model wll all features has gve 39.64% ad 48.36% reducto ukow error relatve to statstcal model respectvely 5. Evaluato I table 4 ad table 5 we llustrate that there are some sgfcat dffereces betwee three models. Sgfcat dffereces are determed by usg pared sample t-test. d* t = sd Follows the t-dstrbuto wth degree of freedom d ( d d) Where d = ad sd = Thus, we obta the followg results show table 4 ad table 5 wth level of sgfcat 0.05.The pared t-test results show the comparso of the statstcal based (ugram, bgram,backoff) techque ad morphologcal edg based (ugram, bgram, backoff) techque uder dfferet expermetal data. These results show that there s sgfcat dfferece amog the three models whe we use these models statstcal techque ad morphologcal edg based techque. I case of backoff model the results are gradually creasg at every pot. Eve the values of ugram ad backoff model have o large dffereces but our t-test result shows that both models ( case of statstcal ad morphologcal edg based techques) are hghly sgfcat. 2 Table 6 shows the average Geometrc mea (G.M) of ukow error relatve to statstcal model. Error dces have bee calculated usg uweghted dex umbers to fd the mprovemet models usg laguage features. G.M as a average gves the average of relatves 6. Coclusos I ths paper, we have proposed combato of two methods for assgg approprate part-of-speech tag to
7 Morphologcal Edg based Strateges of Ukow Word Estmato for Statstcal POS Urdu Tagger 73 word level toke Urdu laguage. Also we have compared two dfferet part-of-speech taggg approaches for Urdu laguage, oe statstcal-based ad the other s statstcal wth morphologcal edg based techque. We have tested ad aalyzed these techques our expermets. Although ths s the frst attempt to apply such approaches for Urdu laguage, yet both of the taggg methods performed well. As a result, our morphologcal edg based approach outperformed purely statstcally based approach uder the gve test codtos. We also ote that due to the reasos of hgh flecto, partally free word order laguage features ad lack of Urdu laguage corpus, purely statstcal based taggg approach dd ot perform well. We have observed that the accuracy of the tagger ca be mproved by addg more features morphologcal edg based model. [3] A. Y. Ng, ad M. I. Jorda, "O Dscrmatve vs. Geeratve Classfers: A Comparso of Logstc Regresso ad Nave Bayes", Advaces Neural Iformato Processg Systems (NIPS), Vol. 4, [4] J.Trommer, ad D. Kallull. "A Morphologcal Tagger for Stadard Albaa", I Proceedgs of LREC,2004 [5] M. Humayou, "Urdu Morphology,Orthography ad lexco Extracto", M.S Thess,Chalmers Uversty of Techology ad Goteborg Uversty, Swede,2006. [6] Y. M. A, H. D. Lm, ad Y. H. Seo, "Korea Part-of-Speech Taggg Based o Cotext Iformato", Idustral Electrocs, Proceedgs. ISIE, 200. Refereces [] S. Brd, E. Kle, ad E. Loper, Itroducto to Natural Laguage Processg, Uversty of Pesylvaa, [2] A. Harde, "Developg a Tagset for Automated Part-of- Speech Taggg Urdu",Corpus Lgustcs coferece. Departmet of Lgustcs, Lacaster Uversty,2003. [3] D. Jurafsky, ad J.H.Mart, A Itroducto to Natural Laguage Processg, Computatoal Lgustcs, ad Speech Recogto, Pretce-Hall, [4] M.Meteer, R.Schwartz, ad R.Weschedel, "Studes Part of Speech Labelg", Huma Laguage Techology Coferece,99. [5] J. Carlbergre, ad V. Ka, "Implemetg a Effcet Part-Of-Speech Tagger", Software: Practce ad Experece, Vol. 29, No. 9, 999, pp [6] C. H. Chag; C. D. Che, "HMM-Based Part-of-Speech Taggg for Chese Corpora", I Proceedgs of ACL-93 Workshop o Very Large Corpora USA, 993. [7] S. Brd, E. Kle, ad E. Loper, Natural Laguage Toolkt (NLTK) [8] Rosefeld, "Scalable Backoff Laguage Models", I Proc. ICSLP'96, Phladelpha, October 996. [9] Z. Dlek, Hakka-Tür, K. Oflazer, G. Tür, "Statstcal Morphologcal Dsambguato for Agglutatve", Computers ad Humates, Vol. 36, 2000, pp [0] C. Oravecz, ad P. Dees, "Effcet Stochastc Part-of- Speech Taggg for Hugara", Proceedgs of the Thrd Iteratoal Coferece o Laguage Resources ad Evaluato, LREC, [] E. Neufeld, ad G. Adams, "Part-of-Speech Taggg from "Small" Data Sets". Ffth Iteratoal Workshop o Artfcal Itellgece ad Statstcs. [2] G. Adams, B. Mllar, E. Neufeld, ad T. Phlp, "Edgbased Strateges for Part-of-Speech Taggg", Ucertaty Artfcal Itellgece(UAI), 994, pp. -7
Clustering documents with vector space model using n-grams
Clusterg documets wth vector space model usg -grams Klas Skogmar, d97ksk@efd.lth.se Joha Olsso, d97jo@efd.lth.se Lud Isttute of Techology Supervsed by: Perre Nugues, Perre.Nugues@cs.lth.se Abstract Ths
More informationOptimal Allocation of Complex Equipment System Maintainability
Optmal Allocato of Complex Equpmet System ataablty X Re Graduate School, Natoal Defese Uversty, Bejg, 100091, Cha edcal Protecto Laboratory, Naval edcal Research Isttute, Shagha, 200433, Cha Emal:rexs841013@163.com
More informationCS 2710 Foundations of AI Lecture 22. Machine learning. Machine Learning
CS 7 Foudatos of AI Lecture Mache learg Mlos Hauskrecht mlos@cs.ptt.edu 539 Seott Square Mache Learg The feld of mache learg studes the desg of computer programs (agets) capable of learg from past eperece
More informationMachine Learning: Algorithms and Applications
/03/ Mache Learg: Algorthms ad Applcatos Florao Z Free Uversty of Boze-Bolzao Faculty of Computer Scece Academc Year 0-0 Lecture 3: th March 0 Naïve Bayes classfer ( Problem defto A trag set X, where each
More informationPoint Estimation-III: General Methods for Obtaining Estimators
Pot Estmato-III: Geeral Methods for Obtag Estmators RECAP 0.-0.6 Data: Radom Sample from a Populato of terest o Real valued measuremets: o Assumpto (Hopefully Reasoable) o Model: Specfed Probablty Dstrbuto
More informationBezier curves. 1. Defining a Bezier curve. A closed Bezier curve can simply be generated by closing its characteristic polygon
Curve represetato Copyrght@, YZU Optmal Desg Laboratory. All rghts reserved. Last updated: Yeh-Lag Hsu (--). Note: Ths s the course materal for ME55 Geometrc modelg ad computer graphcs, Yua Ze Uversty.
More information1-D matrix method. U 4 transmitted. incident U 2. reflected U 1 U 5 U 3 L 2 L 3 L 4. EE 439 matrix method 1
-D matrx method We ca expad the smple plae-wave scatterg for -D examples that we ve see to a more versatle matrx approach that ca be used to hadle may terestg -D problems. The basc dea s that we ca break
More informationFace Recognition using Supervised & Unsupervised Techniques
Natoal Uversty of Sgapore EE5907-Patter recogto-2 NAIONAL UNIVERSIY OF SINGAPORE EE5907 Patter Recogto Project Part-2 Face Recogto usg Supervsed & Usupervsed echques SUBMIED BY: SUDEN NAME: harapa Reddy
More informationArea and Power Efficient Modulo 2^n+1 Multiplier
Iteratoal Joural of Moder Egeerg Research (IJMER) www.jmer.com Vol.3, Issue.3, May-Jue. 013 pp-137-1376 ISSN: 49-6645 Area ad Power Effcet Modulo ^+1 Multpler K. Ptambar Patra, 1 Saket Shrvastava, Sehlata
More informationBlind Steganalysis for Digital Images using Support Vector Machine Method
06 Iteratoal Symposum o Electrocs ad Smart Devces (ISESD) November 9-30, 06 Bld Stegaalyss for Dgtal Images usg Support Vector Mache Method Marcelus Hery Meor School of Electrcal Egeerg ad Iformatcs Badug
More informationDescriptive Statistics: Measures of Center
Secto 2.3 Descrptve Statstcs: Measures of Ceter Frequec dstrbutos are helpful provdg formato about categorcal data, but wth umercal data we ma wat more formato. Statstc: s a umercal measure calculated
More informationNEURO FUZZY MODELING OF CONTROL SYSTEMS
NEURO FUZZY MODELING OF CONTROL SYSTEMS Efré Gorrosteta, Carlos Pedraza Cetro de Igeería y Desarrollo Idustral CIDESI, Av Pe de La Cuesta 70. Des. Sa Pablo. Querétaro, Qro, Méxco gorrosteta@teso.mx pedraza@cdes.mx
More informationDifferentiated Service of Streaming Media Playback Technology
Iteratoal Coferece o Advaced Iformato ad Commucato Techology for Educato (ICAICTE 2013) Dfferetated Servce of Streamg Meda Playback Techology CHENG Z-ao 1 MENG Bo 1 WANG Da-hua 1 ZHAO Yue 1 1 Iformatzato
More informationAPPLICATION OF CLUSTERING METHODS IN BANK S PROPENSITY MODEL
APPLICATION OF CLUSTERING METHODS IN BANK S PROPENSITY MODEL Sergej Srota Haa Řezaková Abstract Bak s propesty models are beg developed for busess support. They should help to choose clets wth a hgher
More informationText Categorization Based on a Similarity Approach
Text Categorzato Based o a Smlarty Approach Cha Yag Ju We School of Computer Scece & Egeerg, Uversty of Electroc Scece ad Techology of Cha, Chegdu 60054, P.R. Cha Abstract Text classfcato ca effcetly ehace
More informationProcess Quality Evaluation based on Maximum Entropy Principle. Yuhong Wang, Chuanliang Zhang, Wei Dai a and Yu Zhao
Appled Mechacs ad Materals Submtted: 204-08-26 ISSN: 662-7482, Vols. 668-669, pp 625-628 Accepted: 204-09-02 do:0.4028/www.scetfc.et/amm.668-669.625 Ole: 204-0-08 204 Tras Tech Publcatos, Swtzerlad Process
More informationITEM ToolKit Technical Support Notes
ITEM ToolKt Notes Fault Tree Mathematcs Revew, Ic. 2875 Mchelle Drve Sute 300 Irve, CA 92606 Phoe: +1.240.297.4442 Fax: +1.240.297.4429 http://www.itemsoft.com Page 1 of 15 6/1/2016 Copyrght, Ic., All
More informationA Comparison of Heuristics for Scheduling Spatial Clusters to Reduce I/O Cost in Spatial Join Processing
Edth Cowa Uversty Research Ole ECU Publcatos Pre. 20 2006 A Comparso of Heurstcs for Schedulg Spatal Clusters to Reduce I/O Cost Spatal Jo Processg Jta Xao Edth Cowa Uversty 0.09/ICMLC.2006.258779 Ths
More informationJournal of Chemical and Pharmaceutical Research, 2015, 7(3): Research Article
Avalable ole www.ocpr.com Joural of Chemcal ad Pharmaceutcal Research, 2015, 73):476-481 Research Artcle ISSN : 0975-7384 CODENUSA) : JCPRC5 Research o cocept smlarty calculato method based o sematc grd
More informationChapter 3 Descriptive Statistics Numerical Summaries
Secto 3.1 Chapter 3 Descrptve Statstcs umercal Summares Measures of Cetral Tedecy 1. Mea (Also called the Arthmetc Mea) The mea of a data set s the sum of the observatos dvded by the umber of observatos.
More informationWeighting Cache Replace Algorithm for Storage System
Weghtg Cache Replace Algorthm for Storage System Yhu Luo 2 Chagsheg Xe 2 Chegfeg Zhag 2 School of mathematcs ad Computer Scece, Hube Uversty, Wuha 430062, P.R. Cha 2 Natoal Storage System Laboratory, School
More informationAN IMPROVED TEXT CLASSIFICATION METHOD BASED ON GINI INDEX
Joural of Theoretcal ad Appled Iformato Techology 30 th September 0. Vol. 43 No. 005-0 JATIT & LLS. All rghts reserved. ISSN: 99-8645 www.jatt.org E-ISSN: 87-395 AN IMPROVED TEXT CLASSIFICATION METHOD
More informationEstimation of Co-efficient of Variation in PPS sampling.
It. Statstcal Ist.: Proc. 58th World Statstcal Cogress, 0, Dubl (Sesso CPS00) p.409 Estmato of Co-effcet of Varato PPS samplg. Archaa. V ( st Author) Departmet of Statstcs, Magalore Uverst Magalagagotr,
More informationAPR 1965 Aggregation Methodology
Sa Joaqu Valley Ar Polluto Cotrol Dstrct APR 1965 Aggregato Methodology Approved By: Sged Date: March 3, 2016 Araud Marjollet, Drector of Permt Servces Backgroud Health rsk modelg ad the collecto of emssos
More informationConstructive Semi-Supervised Classification Algorithm and Its Implement in Data Mining
Costructve Sem-Supervsed Classfcato Algorthm ad Its Implemet Data Mg Arvd Sgh Chadel, Arua Twar, ad Naredra S. Chaudhar Departmet of Computer Egg. Shr GS Ist of Tech.& Sc. SGSITS, 3, Par Road, Idore (M.P.)
More informationUsing The ACO Algorithm in Image Segmentation for Optimal Thresholding 陳香伶財務金融系
教專研 95P- Usg The ACO Algorthm Image Segmetato for Optmal Thresholdg Abstract Usg The ACO Algorthm Image Segmetato for Optmal Thresholdg 陳香伶財務金融系 Despte the fact that the problem of thresholdg has bee qute
More informationANALYSIS OF VARIANCE WITH PARETO DATA
Proceedgs of the th Aual Coferece of Asa Pacfc Decso Sceces Isttute Hog Kog, Jue -8, 006, pp. 599-609. ANALYSIS OF VARIANCE WITH PARETO DATA Lakhaa Watthaacheewakul Departmet of Mathematcs ad Statstcs,
More informationPerformance Impact of Load Balancers on Server Farms
erformace Impact of Load Balacers o Server Farms Ypg Dg BMC Software Server Farms have gaed popularty for provdg scalable ad relable computg / Web servces. A load balacer plays a key role ths archtecture,
More informationNew Fuzzy Integral for the Unit Maneuver in RTS Game
New Fuzzy Itegral for the Ut Maeuver RTS Game Peter Hu Fug Ng, YgJe L, ad Smo Ch Keug Shu Departmet of Computg, The Hog Kog Polytechc Uversty, Hog Kog {cshfg,csyjl,csckshu}@comp.polyu.edu.hk Abstract.
More informationUnsupervised Discretization Using Kernel Density Estimation
Usupervsed Dscretzato Usg Kerel Desty Estmato Maregle Bba, Floraa Esposto, Stefao Ferll, Ncola D Mauro, Teresa M.A Basle Departmet of Computer Scece, Uversty of Bar Va Oraboa 4, 7025 Bar, Italy {bba,esposto,ferll,dm,basle}@d.uba.t
More informationUsing Linear-threshold Algorithms to Combine Multi-class Sub-experts
Usg Lear-threshold Algorthms to Combe Mult-class Sub-experts Chrs Mesterharm MESTERHA@CS.RUTGERS.EDU Rutgers Computer Scece Departmet 110 Frelghuyse Road Pscataway, NJ 08854 USA Abstract We preset a ew
More informationPreventing Information Leakage in C Applications Using RBAC-Based Model
Proceedgs of the 5th WSEAS It. Cof. o Software Egeerg Parallel ad Dstrbuted Systems Madrd Spa February 5-7 2006 (pp69-73) Prevetg Iformato Leakage C Applcatos Usg RBAC-Based Model SHIH-CHIEN CHOU Departmet
More informationCLUSTERING ASSISTED FUNDAMENTAL MATRIX ESTIMATION
CLUSERING ASSISED FUNDAMENAL MARIX ESIMAION Hao Wu ad Y Wa School of Iformato Scece ad Egeerg, Lazhou Uversty, Cha wuhao1195@163com, wayjs@163com ABSRAC I computer vso, the estmato of the fudametal matrx
More informationTDT-2004: ADAPTIVE TOPIC TRACKING AT MARYLAND
TDT-2004: ADAPTIVE TOPIC TRACKING AT MARYLAND Tamer Elsayed, Douglas W. Oard, Davd Doerma Isttute for Advaced r Studes Uversty of Marylad, College Park, MD 20742 Cotact author: telsayed@cs.umd.edu Gary
More informationA Genetic K-means Clustering Algorithm Applied to Gene Expression Data
A Geetc K-meas Clusterg Algorthm Appled to Gee Expresso Data Fag-Xag Wu, W. J. Zhag, ad Athoy J. Kusal Dvso of Bomedcal Egeerg, Uversty of Sasatchewa, Sasatoo, S S7N 5A9, CANADA faw34@mal.usas.ca, zhagc@egr.usas.ca
More informationDenoising Algorithm Using Adaptive Block Based Singular Value Decomposition Filtering
Advaces Computer Scece Deosg Algorthm Usg Adaptve Block Based Sgular Value Decomposto Flterg SOMKAIT UDOMHUNSAKUL Departmet of Egeerg ad Archtecture Rajamagala Uversty of Techology Suvarabhum 7/ Suaya,
More informationSVM Classification Method Based Marginal Points of Representative Sample Sets
P P College P P College P Iteratoal Joural of Iformato Techology Vol. No. 9 005 SVM Classfcato Method Based Margal Pots of Represetatve Sample Sets Wecag ZhaoP P, Guagrog JP P, Ru NaP P, ad Che FegP of
More informationIdentifying Chemical Names in Biomedical Text: An Investigation of the Substring Co-occurrence Based Approaches
Idetfyg Chemcal Names Bomedcal Text: A Ivestgato of the Substrg Co-occurrece Based Approaches Alexader Vasserma Departmet of Computer ad Iformato Scece Uversty of Pesylvaa Phladelpha, PA 94 avasserm@seas.upe.edu
More informationA New Hybrid Audio Classification Algorithm Based on SVM Weight Factor and Euclidean Distance
Proceedgs of the 2007 WSEAS Iteratoal Coferece o Computer Egeerg ad Applcatos, Gold Coast, Australa, Jauary 7-9, 2007 52 A New Hybrd Audo Classfcato Algorthm Based o SVM Weght Factor ad Eucldea Dstace
More informationSoftware Clustering Techniques and the Use of Combined Algorithm
Software Clusterg Techques ad the Use of Combed Algorthm M. Saeed, O. Maqbool, H.A. Babr, S.Z. Hassa, S.M. Sarwar Computer Scece Departmet Lahore Uversty of Maagemet Sceces DHA Lahore, Paksta oaza@lums.edu.pk
More information2 General Regression Neural Network (GRNN)
4 Geeral Regresso Neural Network (GRNN) GRNN, as proposed b oald F. Specht [Specht 9] falls to the categor of probablstc eural etworks as dscussed Chapter oe. Ths eural etwork lke other probablstc eural
More informationA hybrid method using FAHP and TOPSIS for project selection Xuan Lia, Jiang Jiangb and Su Deng c
5th Iteratoal Coferece o Computer Sceces ad Automato Egeerg (ICCSAE 205) A hybrd method usg FAHP ad TOPSIS for project selecto Xua La, Jag Jagb ad Su Deg c College of Iformato System ad Maagemet, Natoal
More informationApplication of Genetic Algorithm for Computing a Global 3D Scene Exploration
Joural of Software Egeerg ad Applcatos, 2011, 4, 253-258 do:10.4236/jsea.2011.44028 Publshed Ole Aprl 2011 (http://www.scrp.org/joural/jsea) 253 Applcato of Geetc Algorthm for Computg a Global 3D Scee
More informationIntegrating Language Model in Handwritten Chinese Text Recognition
2009 10th Iteratoal Coferece o Documet Aalyss ad Recogto Itegratg Laguage Model Hadwrtte Chese Text Recogto Qu-Feg Wag, Fe Y, Cheg-L Lu Natoal Laboratory of Patter Recogto (NLPR), Isttute of Automato,
More informationFuzzy ID3 Decision Tree Approach for Network Reliability Estimation
IJCSI Iteratoal Joural of Computer Scece Issues, Vol. 9, Issue 1, o 1, Jauary 2012 ISS (Ole): 1694-0814 www.ijcsi.org 446 Fuzzy ID3 Decso Tree Approach for etwor Relablty Estmato A. Ashaumar Sgh 1, Momtaz
More informationA Double-Window-based Classification Algorithm for Concept Drifting Data Streams
00 IEEE Iteratoal Coferece o Graular Computg A Double-Wdow-based Classfcato Algorthm for Cocept Drftg Data Streams Qu Zhu, Xuegag Hu, Yuhog Zhag, Pepe L, Xdog Wu, School of Computer Scece ad Iformato Egeerg,
More informationEstimating Feasibility Using Multiple Surrogates and ROC Curves
Estmatg Feasblty Usg Multple Surrogates ad ROC Curves Arba Chaudhur * Uversty of Florda, Gaesvlle, Florda, 3601 Rodolphe Le Rche École Natoale Supéreure des Mes de Sat-Étee, Sat-Étee, Frace ad CNRS LIMOS
More informationAn Optimized Algorithm for Big Data Classification using Neuro Fuzzy Approach
Ida Joural of Scece ad Techology, Vol 9(28), DOI: 0.7485/jst/206/v928/87995, July 206 ISSN (Prt) : 0974-6846 ISSN (Ole) : 0974-5645 A Optmzed Algorthm for Bg Data Classfcato usg Neuro Fuzzy Approach Naveet
More informationCollaborative Filtering Support for Adaptive Hypermedia
Collaboratve Flterg Support for Adaptve Hypermeda Mart Balík, Iva Jelíek Departmet of Computer Scece ad Egeerg, Faculty of Electrcal Egeerg Czech Techcal Uversty Karlovo áměstí 3, 35 Prague, Czech Republc
More informationVertex Odd Divisor Cordial Labeling of Graphs
IJISET - Iteratoal Joural of Iovatve Scece, Egeerg & Techology, Vol. Issue 0, October 0. www.jset.com Vertex Odd Dvsor Cordal Labelg of Graphs ISSN 48 68 A. Muthaya ad P. Pugaleth Assstat Professor, P.G.
More informationAn Ensemble Multi-Label Feature Selection Algorithm Based on Information Entropy
The Iteratoal Arab Joural of Iformato Techology, Vol., No. 4, July 204 379 A Esemble Mult-Label Feature Selecto Algorthm Based o Iformato Etropy Shg L, Zheha Zhag, ad Jaq Dua School of Computer Scece,
More informationAutomated approach for the surface profile measurement of moving objects based on PSP
Uversty of Wollogog Research Ole Faculty of Egeerg ad Iformato Sceces - Papers: Part B Faculty of Egeerg ad Iformato Sceces 207 Automated approach for the surface profle measuremet of movg objects based
More informationPriority-based Packet Scheduling in Internet Protocol Television
EMERGING 0 : The Thrd Iteratoal Coferece o Emergg Network Itellgece Prorty-based Packet Schedulg Iteret Protocol Televso Mehmet Dez Demrc Computer Scece Departmet Istabul Uversty İstabul, Turkey e-mal:demrcd@stabul.edu.tr
More informationChEn 475 Statistical Analysis of Regression Lesson 1. The Need for Statistical Analysis of Regression
Statstcal-Regresso_hadout.xmcd Statstcal Aalss of Regresso ChE 475 Statstcal Aalss of Regresso Lesso. The Need for Statstcal Aalss of Regresso What do ou do wth dvdual expermetal data pots? How are the
More informationFEATURE SELECTION ON COMBINATIONS FOR EFFICIENT LEARNING FROM IMAGES. Rong Xiao, Lei Zhang, and Hong-Jiang Zhang
FEATURE SELECTION ON COMBINATIONS FOR EFFICIENT LEARNING FROM IMAGES Rog Xao, Le Zhag, ad Hog-Jag Zhag Mcrosoft Research Asa, Bejg 100080, P.R. Cha {t-rxao, lezhag, hjzhag}@mcrosoft.com ABSTRACT Due to
More informationSpeeding- Up Fractal Image Compression Using Entropy Technique
Avalable Ole at www.jcsmc.com Iteratoal Joural of Computer Scece ad Moble Computg A Mothly Joural of Computer Scece ad Iformato Techology ISSN 2320 088X IMPACT FACTOR: 5.258 IJCSMC, Vol. 5, Issue. 4, Aprl
More informationCOMSC 2613 Summer 2000
Programmg II Fal Exam COMSC 63 Summer Istructos: Name:. Prt your ame the space provded Studet Id:. Prt your studet detfer the space Secto: provded. Date: 3. Prt the secto umber of the secto whch you are
More informationBeijing University of Technology, Beijing , China; Beijing University of Technology, Beijing , China;
d Iteratoal Coferece o Machery, Materals Egeerg, Chemcal Egeerg ad Botechology (MMECEB 5) Research of error detecto ad compesato of CNC mache tools based o laser terferometer Yuemg Zhag, a, Xuxu Chu, b
More informationSoftware reliability is defined as the probability of failure
Evolutoary Regresso Predcto for Software Cumulatve Falure Modelg: a comparatve study M. Beaddy, M. Wakrm & S. Aljahdal 2 : Dept. of Math. & Ifo. Equpe MMS, Ib Zohr Uversty Morocco. beaddym@yahoo.fr 2:
More informationA Comparison of Univariate Smoothing Models: Application to Heart Rate Data Marcus Beal, Member, IEEE
A Comparso of Uvarate Smoothg Models: Applcato to Heart Rate Data Marcus Beal, Member, IEEE E-mal: bealm@pdx.edu Abstract There are a umber of uvarate smoothg models that ca be appled to a varety of olear
More informationImpact of Mobility Prediction on the Temporal Stability of MANET Clustering Algorithms *
Impact of Moblty Predcto o the Temporal Stablty of MANET Clusterg Algorthms * Aravdha Vekateswara, Vekatesh Saraga, Nataraa Gautam 1, Ra Acharya Departmet of Comp. Sc. & Egr. Pesylvaa State Uversty Uversty
More informationSpatial Interpolation Using Neural Fuzzy Technique
Wog, K.W., Gedeo, T., Fug, C.C. ad Wog, P.M. (00) Spatal terpolato usg eural fuzzy techque. I: Proceedgs of the 8th Iteratoal Coferece o Neural Iformato Processg (ICONIP), Shagha, Cha Spatal Iterpolato
More informationEvaluation of Node and Link Importance Based on Network Topology and Traffic Information DU Xun-Wei, LIU Jun, GUO Wei
Advaced Materals Research Submtted: 2014-08-29 ISSN: 1662-8985, Vols. 1049-1050, pp 1765-1770 Accepted: 2014-09-01 do:10.4028/www.scetfc.et/amr.1049-1050.1765 Ole: 2014-10-10 2014 Tras Tech Publcatos,
More informationAdaptive Clustering Algorithm for Mining Subspace Clusters in High-Dimensional Data Stream *
ISSN 673-948 CODEN JKYTA8 E-mal: fcst@vp.63.com Joural of Froters of Computer Scece ad Techology http://www.ceaj.org 673-948/200/04(09)-0859-06 Tel: +86-0-566056 DOI: 0.3778/j.ss.673-948.200.09.009 *,2,
More informationInteractive Change Detection Using High Resolution Remote Sensing Images Based on Active Learning with Gaussian Processes
Iteractve Chage Detecto Usg Hgh Resoluto Remote Sesg Images Based o Actve Learg wth Gaussa Processes Hu Ru a, Hua Yu a,, Pgpg Huag b, We Yag a a School of Electroc Iformato, Wuha Uversty, 43007 Wuha, Cha
More informationActive Bayesian Learning For Mixture Models
Actve Bayesa Learg For Mxture Models Ia Davdso Slco Graphcs 300 Crttede L, MS 876 Mouta Vew, CA 94587 pd@hotmal.com Abstract Tradtoally, Bayesa ductve learg volves fdg the most probable model from the
More informationDeveloper Recommendation with Awareness of Accuracy and Cost
Developer Recommedato wth Awareess of Accuracy ad Cost * J Lu, Yquz Ta State Key Lab of Software Egeerg Computer School, Wuha Uversty Wuha, Cha *Correspodg author jlu@whu.edu.c tayquz@whu.edu.c Lag Hog
More informationA Web Mining Based Network Personalized Learning System Hua PANG1, a, Jian YU1, Long WANG2, b
3rd Iteratoal Coferece o Machery, Materals ad Iformato Techology Applcatos (ICMMITA 05) A Web Mg Based Network Persoalzed Learg System Hua PANG, a, Ja YU, Log WANG, b College of Educato Techology, Sheyag
More informationPERSPECTIVES OF THE USE OF GENETIC ALGORITHMS IN CRYPTANALYSIS
PERSPECTIVES OF THE USE OF GENETIC ALGORITHMS IN CRYPTANALYSIS Lal Besela Sokhum State Uversty, Poltkovskaa str., Tbls, Georga Abstract Moder cryptosystems aalyss s a complex task, the soluto of whch s
More informationKeywords- clustering; naïve Bayesian classifier; boosting; hybrid classifier.
World of Computer Scece ad Iformato Techology Joural (WCSIT) ISSN: 2221-0741 Vol. 1, No. 3, 105-109, 2011. A Hybrd Classfer usg Boostg, Clusterg, ad Naïve Bayesa Classfer A. J. M. Abu Afza, Dewa Md. Fard,
More informationReconstruction of Orthogonal Polygonal Lines
Recostructo of Orthogoal Polygoal Les Alexader Grbov ad Eugee Bodasky Evrometal System Research Isttute (ESRI) 380 New ork St. Redlads CA 9373-800 USA {agrbov ebodasky}@esr.com Abstract. A orthogoal polygoal
More informationClassification Web Pages By Using User Web Navigation Matrix By Mementic Algorithm
Classfcato Web Pages By Usg User Web Navgato Matrx By Memetc Algorthm 1 Parvaeh roustae 2 Mehd sadegh zadeh 1 Studet of Computer Egeerg Software EgeergDepartmet of ComputerEgeerg, Bushehr brach,
More informationFor all questions, answer choice E) NOTA" means none of the above answers is correct. A) 50,500 B) 500,000 C) 500,500 D) 1,001,000 E) NOTA
For all questos, aswer choce " meas oe of the above aswers s correct.. What s the sum of the frst 000 postve tegers? A) 50,500 B) 500,000 C) 500,500 D),00,000. What s the sum of the tegers betwee 00 ad
More informationMulticlass classification Decision trees
CS 75 Mache Learg Lecture Multclass classfcato Decso trees Mlos Hauskrecht mlos@cs.tt.edu 59 Seott Suare CS 75 Mache Learg Mdterm eam Mdterm Tuesda, March 4, 4 I-class 75 mutes closed book materal covered
More informationOptimization of Light Switching Pattern on Large Scale using Genetic Algorithm
Optmzato of Lght Swtchg Patter o Large Scale usg Geetc Algorthm Pryaka Sambyal, Pawaesh Abrol 2, Parvee Lehaa 3,2 Departmet of Computer Scece & IT 3 Departmet of Electrocs Uversty of Jammu, Jammu, J&K,
More informationEDGE- ODD Gracefulness of the Tripartite Graph
EDGE- ODD Graceuless o the Trpartte Graph C. Vmala, A. Saskala, K. Ruba 3, Asso. Pro, Departmet o Mathematcs, Peryar Maamma Uversty, Vallam, Thajavur Post.. Taml Nadu, Ida. 3 M. Phl Scholar, Departmet
More informationComparison Studies on Classification for Remote Sensing Image Based on Data Mining Method
Hag Xao ad Xub Zhag Comparso Studes o Classfcato for Remote Sesg Image Based o Data Mg Method Hag XIAO 1, Xub ZHANG 1 1: School of Electroc, Iformato ad Electrcal Egeerg Shagha Jaotog Uversty No. 1954,
More informationContent-Based Image Retrieval Using Associative Memories
Proceedgs of the 6th WSEAS It. Coferece o ELECOMMUNICAIONS ad INFORMAICS, Dallas, exas, USA, March 22-24, 2007 99 Cotet-Based Image Retreval Usg Assocatve Memores ARUN KULKARNI Computer Scece Departmet
More informationMarcus Gallagher School of Information Technology and Electrical Engineering The University of Queensland QLD 4072, Australia
O the Importace of Dversty Mateace Estmato of Dstrbuto Algorthms Bo Yua School of Iformato Techology ad Electrcal Egeerg The Uversty of Queeslad QLD 4072, Australa +6-7-3365636 boyua@tee.uq.edu.au Marcus
More informationFingerprint Classification Based on Spectral Features
Fgerprt Classfcato Based o Spectral Features Hosse Pourghassem Tarbat Modares Uversty h_poorghasem@modares.ac.r Hassa Ghassema Tarbat Modares Uversty ghassem@modares.ac.r Abstract: Fgerprt s oe of the
More informationDelay based Duplicate Transmission Avoid (DDA) Coordination Scheme for Opportunistic routing
Delay based Duplcate Trasmsso Avod (DDA) Coordato Scheme for Opportustc routg Ng L, Studet Member IEEE, Jose-Fera Martez-Ortega, Vcete Heradez Daz Abstract-Sce the packet s trasmtted to a set of relayg
More informationNetwork Security Evaluation Based on Variable Weight Fuzzy Cloud Model
207 2 d Iteratoal Coferece o Computer Scece ad Techology (CST 207) ISBN: 978--60595-46-5 Networ Securty Evaluato Based o Varable Weght Fuzzy Cloud Model Yag JIANG a*, Cheg-ha LI, Zh-peg LI ad Mg-ca SUN
More informationMachine Learning. CS 2750 Machine Learning. Administration. Lecture 1. Milos Hauskrecht 5329 Sennott Square, x4-8845
CS 75 Mache Learg Lecture Mache Learg Mlos Hauskrecht mlos@cs.ptt.edu 539 Seott Square, 5 people.cs.ptt.edu/~mlos/courses/cs75/ Admstrato Istructor: Prof. Mlos Hauskrecht mlos@cs.ptt.edu 539 Seott Square,
More informationEnumerating XML Data for Dynamic Updating
Eumeratg XML Data for Dyamc Updatg Lau Ho Kt ad Vcet Ng Departmet of Computg, The Hog Kog Polytechc Uversty, Hug Hom, Kowloo, Hog Kog cstyg@comp.polyu.edu.h Abstract I ths paper, a ew mappg model, called
More informationFace Authentication for Multiple Subjects Using Eigenflow
Face Authetcato for Multple Subjects Usg Egeflow Xaomg Lu Tsuha Che ad B.V.K. Vjaya Kumar Advaced Multmeda Processg Lab Techcal Report AMP -5 Aprl 2 Electrcal ad Computer Egeerg Carege Mello Uversty Pttsburgh,
More informationApplying Support Vector Machines to Imbalanced Datasets
Applyg Support Vector Maches to Imbalaced Datasets Reha Akba 1, Stephe Kwek 1, ad Nathale Japkowcz 2 1 Departmet of Computer Scece, Uversty of Texas at Sa Atoo 6900 N. Loop 1604 W, Sa Atoo, Texas, 78249,
More informationParallel Iterative Poisson Solver for a Distributed Memory Architecture
Parallel Iteratve Posso Solver for a Dstrbted Memory Archtectre Erc Dow Aerospace Comptatoal Desg Lab Departmet of Aeroatcs ad Astroatcs 2 Motvato Solvg Posso s Eqato s a commo sbproblem may mercal schemes
More informationMulti-dimensional Characteristics Analysis of Decathlon Champion Achievement of Modern Olympic Game
Iteratoal Workshop o Computer Scece Sports (IWCSS 203) Mult-dmesoal Characterstcs Aalyss of Decathlo Champo Achevemet of Moder Olympc Game Chagme Huag Isttute of Physcal Educato Hua Uversty of Scece ad
More informationA genetic procedure used to train RFB neural networks
A geetc procedure used to tra RFB eural etworks Costat-Iula VIZITIU Commucatos ad Electroc Systems Departmet Mltary Techcal Academy George Cosbuc Aveue 8-83 5 th Dstrct Bucharest ROMANIA vc@mta.ro http://www.mta.ro
More informationWEB PAGE CLASSIFIERS FOR TOPICAL CRAWLER
244 WEB PAGE CLASSIFIERS FOR TOPICAL CRAWLER Dw H. Wdyatoro, Masayu L. Khodra, Paramta * School of Iformatcs ad Electrcal Egeerg - Isttut Tekolog Badug Jl. Gaesha 0 Badug Telephoe 62 22 250835 emal : dw@f.tb.ac.d,
More informationA New Newton s Method with Diagonal Jacobian Approximation for Systems of Nonlinear Equations
Joural of Mathematcs ad Statstcs 6 (3): 46-5, ISSN 549-3644 Scece Publcatos A New Newto s Method wth Dagoal Jacoba Appromato for Systems of Nolear Equatos M.Y. Wazr, W.J. Leog, M.A. Hassa ad M. Mos Departmet
More informationMode-based temporal filtering for in-band wavelet video coding with spatial scalability
Mode-based temporal flterg for -bad wavelet vdeo codg wth spatal scalablty ogdog Zhag a*, Jzheg Xu b, Feg Wu b, Weju Zhag a, ogka Xog a a Image Commucato Isttute, Shagha Jao Tog Uversty, Shagha b Mcrosoft
More informationPersonalized Search Based on Context-Centric Model
JOURNAL OF NETWORKS, VOL. 8, NO. 7, JULY 13 1 Persoalzed Search Based o Cotext-Cetrc Model Mgyag Lu, Shufe Lu, Chaghog Hu Dept. College of Computer Scece ad Techology, Jl Uversty, Chagchu, Jl, Cha Emal:
More informationAn Ensemble Approach to Classifier Construction based on Bootstrap Aggregation
A Esemble Approach to Classfer Costructo based o Bootstrap Aggregato Dewa Md. Fard Jahagragar Uversty Dhaka-342, Bagladesh Mohammad Zahdur Rahma Jahagragar Uversty Dhaka-342, Bagladesh Chowdhury Mofzur
More informationWeb Page Clustering by Combining Dense Units
Web Page Clusterg by Combg Dese Uts Morteza Haghr Chehregha, Hassa Abolhassa ad Mostafa Haghr Chehregha Departmet of CE, Sharf Uversty of Techology, Tehra, IRA {haghr, abolhassa}@ce.sharf.edu Departmet
More informationA new approach based in mean and standard deviation for authentication system of face
M. Fedas, D. Sagaa A ew approach based mea ad stadard devato for authetcato system of face M. Fedas 1, D. Sagaa 2 Abstract Face authetcato s a sgfcat problem patter recogto. The face s ot rgd t ca udergo
More informationInternational Mathematical Forum, 1, 2006, no. 31, ON JONES POLYNOMIALS OF GRAPHS OF TORUS KNOTS K (2, q ) Tamer UGUR, Abdullah KOPUZLU
Iteratoal Mathematcal Forum,, 6, o., 57-54 ON JONES POLYNOMIALS OF RAPHS OF TORUS KNOTS K (, q ) Tamer UUR, Abdullah KOPUZLU Atatürk Uverst Scece Facult Dept. of. Math. 54 Erzurum, Turkey tugur@atau.edu.tr
More informationAnalysis of Students' Performance by Using Different Data Mining Classifiers
I.J. Moder Educato ad Computer Scece, 2017, 8, 9-15 Publshed Ole August 2017 MECS (http://www.mecs-press.org/) DOI: 10.5815/jmecs.2017.08.02 Aalyss of Studets' Performace by Usg Dfferet Data Mg Classfers
More informationMATHEMATICAL PROGRAMMING MODEL OF THE CRITICAL CHAIN METHOD
MATHEMATICAL PROGRAMMING MODEL OF THE CRITICAL CHAIN METHOD TOMÁŠ ŠUBRT, PAVLÍNA LANGROVÁ CUA, SLOVAKIA Abstract Curretly there s creasgly dcated that most of classcal project maagemet methods s ot sutable
More informationTwo step approach for Software Process Control: HLSRGM
Iteratoal Joural of Emergg Treds & Techology Computer Scece (IJETTCS Web Ste: wwwjettcsorg Emal: edtor@jettcsorg, edtorjettcs@gmalcom Volume, Issue 4, July August 03 ISS 78-686 Two step approach for Software
More information