Learning Semantics-Preserving Distance Metrics for Clustering Graphical Data

Size: px
Start display at page:

Download "Learning Semantics-Preserving Distance Metrics for Clustering Graphical Data"

Transcription

1 Learnng Semantcs-Preservng Dstance Metrcs for Clusterng Graphcal Data Aparna S. Varde, Elke A. Rundenstener Carolna Ruz Mohammed Manruzzaman,3 Rchard D. Ssson Jr.,3 Department of Computer Scence Center for Heat Treatng Excellence 3 Department of Mechancal Engneerng Worcester Polytechnc Insttute (WPI) Worcester, MA USA (aparna rundenst ruz manruzz ssson)@wp.edu ABSTRACT In mnng graphcal data the default Eucldean dstance s often used as a noton of smlarty. However ths does not adequately capture semantcs n our targeted domans, havng graphcal representatons depctng results of scentfc experments. It s seldom known a-pror what other dstance metrc best preserves semantcs. Ths motvates the need to learn such a metrc. A technque called LearnMet s proposed here to learn a domanspecfc dstance metrc for graphcal representatons. Input to LearnMet s a tranng set of correct clusters of such graphs. LearnMet teratvely compares these correct clusters wth those obtaned from an arbtrary but fxed clusterng algorthm. In the frst teraton a guessed metrc s used for clusterng. Ths metrc s then refned usng the error between the obtaned and correct clusters untl the error s below a gven threshold. LearnMet s evaluated rgorously n the Heat Treatng doman whch motvated ths research. Clusters obtaned usng the learned metrc and clusters obtaned usng Eucldean dstance are both compared aganst the correct clusters over a separate test set. Our results show that the learned metrc provdes better clusters. Categores and Subject Descrptors I..6 [Artfcal Intellgence]: Learnng Parameter learnng. General Terms Algorthms, Expermentaton. Keywords Dstance metrc, clusterng, semantc graphcal mnng.. INTRODUCTION Expermental results n scentfc domans are often represented graphcally. Such graphs depct the functonal behavor of process parameters hence ncorporatng semantcs. In ths paper, we use the term graph to mean such a graphcal representaton. In mnng graphs wth technques such as clusterng [9] the measure for comparson s typcally Eucldean dstance [see e.g. 7] whch often poses problems. For example, n the doman of Heat Treatng of Materals whch motvated ths work, graphs are used to plot the heat transfer versus temperature durng the rapd coolng of a materal [see e.g. 3]. Crtcal regons on these graphs represent sgnfcant physcal phenomena n the doman. Any algorthm (based for nstance on Eucldean dstance) that consders two graphs as smlar (relatve to other graphs) although ther crtcal regons dffer s consdered semantcally ncorrect [3,4]. Lkewse n several scentfc domans, there could be sgnfcant features on graphs. Knowledge of these features and ther relatve mportance may at best be avalable n a subjectve form, but not as a metrc. Ths motvates the development of a technque to learn dstance metrcs that capture the semantcs of the graphs. Hnneburg et al [6] propose a learnng method to fnd the relatve mportance of dmensons for n-dmensonal objects. However, ther focus s on dmensonalty reducton and not on doman semantcs. In [6] they learn whch type of poston-based dstance s applcable for the gven data startng from the formula of Mahalanobs dstance. However they do not deal wth graphcal data and semantcs. Kem et al. [8] overvew varous dstance types for smlarty search over multmeda databases. However no sngle dstance measure encompassng several types s proposed. Lnear regresson [] and neural networks [] could possbly be used for learnng a doman-specfc dstance metrc for graphs. However these technques do not acheve accuracy acceptable n our targeted domans [4]. Genetc algorthms [5] f used for feature selecton n graphs also gve the problem of nsuffcent accuracy. Ths s due to lack of doman knowledge [4]. Fourer transforms [4] f used to represent the graphs do not preserve the crtcal regons n the doman due to the nature of the transform [4]. Hence they are not accurate enough n capturng semantcs. Accuracy n ths context s measured by evaluatng the effectveness of a gven metrc n mnng unseen data [4]. We propose an approach called LearnMet to learn a dstance metrc for graphs ncorporatng doman semantcs. The nput to LearnMet s a tranng set wth correct (.e., gven by doman experts) clusters of graphs over a subset of the expermental data. The steps of our approach are: () guess ntal metrc guded by doman knowledge; () use that metrc for clusterng wth an arbtrary but fxed clusterng algorthm; (3) evaluate accuracy of obtaned clusters by comparng them wth correct clusters; (4) adjust metrc based on error between obtaned and correct clusters, f error s below threshold or f executon tmes out then termnate and go to step (5), else go to step (); and (5) once

2 termnated, output the metrc gvng error below threshold, or mnmum error so far as the learned metrc. LearnMet s evaluated usng a dstnct test set of correct clusters of graphs provded by experts. The learned metrc s used to cluster the graphs n the test set. The obtaned clusters are compared wth correct clusters n the test set. The closer the obtaned clusters match the correct clusters, the lower the error. The clusters obtaned usng the default noton of Eucldean dstance, are also compared wth the correct clusters. The dfference n the error denotes the effectveness of the learned metrc. The LearnMet metrc consstently outperformed the Eucldean metrc n our experments. LearnMet s also evaluated by ntegratng t wth the AutoDomanMne system desgned for computatonal estmaton of process parameters n scentfc experments [3]. Ths s explaned n Secton 4.. DISTANCE METRICS We descrbe dstance metrc types relevant to our domans. Let A (A,A...A n ) and B (B,B...B n ) be two n-dmensonal objects. Poston-based Dstances: They refer to absolute poston of objects [see e.g. 7]. Examples are: Eucldean Dstance: As-the-crow-fles dstance between objects. D Eucldean = n = ( A B ) Manhattan Dstance: Cty-block dstance between objects. n DManhat tan = A B = Statstcal Dstances: Ths refers to dstances based on statstcal features [see e.g. ]. Examples are: Mean Dstance: Dstance between mean values of the objects: D Mean = Mean( A) Mean( Maxmum Dstance: Dstance between maxmum values: D Max = Max( A) Max( In addton to these, we ntroduce the concept of Crtcal Dstance for graphs as applcable to our targeted domans [4]. Crtcal Dstances: Gven graphs A and B, a Crtcal Dstance s a dstance metrc between crtcal regons of A and B where a crtcal regon represents the occurrence of a sgnfcant physcal phenomenon. They are calculated n a doman-specfc manner. Fgure : Crtcal Dstance Examples One example of crtcal dstance as shown n Fgure s the Ledenfrost dstance [3,4] n Heat Treatng. Ths s for Ledenfrost Ponts [see e.g. 3] whch denote the breakng of the vapor blanket around a part n heat treatment. It s calculated as D LF = ( ATLF BTLF ) + ( AhLF BhLF ) where T LF s the temperature at Ledenfrost Pont and h LF s the heat transfer coeffcent at that pont [0]. Another crtcal dstance s the Bolng Pont dstance D BP [3,4]. Ths s the dstance between ponts on the graphs correspondng to the Bolng Ponts [see e.g. 3] of the respectve coolng meda n heat treatment. D temperature and h BP s heat transfer coeffcent at Bolng Pont. BP = ( ATBP BTBP ) + ( AhBP BhBP ) where T BP s 3. THE LEARNMET STRATEGY To gve detals of LearnMet, a dstance metrc s frst defned. Dstance Metrc n LearnMet: A LearnMet dstance metrc D s a weghted sum of components, where each component can be a poston-based, a statstcal, or a crtcal dstance metrc. The weght of each component s a numercal value ndcatng ts relatve mportance n the doman. Thus a LearnMet dstance metrc s of the form D = w Dc w Dc where each Dc m m s a component, w s ts weght, and m s number of components. As requred n our applcaton doman, D should be a metrc so that clusterng algorthms requrng the noton of smlarty to be a dstance metrc can be used; ndexng structures such as B+ trees for metrcs can be appled; and prunng n smlarty search can be performed usng trangle nequalty. Condtons for D to be a metrc are stated as a theorem below. Theorem : If each component Dc s a dstance metrc and each weght 0 m w then D = = w Dc s a dstance metrc,.e., t satsfes the metrc propertes. The proof of ths theorem s straghtforward and can be found n [4]. In our targeted domans, condtons n Theorem are satsfed. Snce our graphs nvolve nterval-scaled varables and the fundamental dstance types applcable to these are metrcs [see e.g. 7], ths s suffcent to say that each Dc s a metrc. Also, we consder only non-negatve weghts. The LearnMet steps are dscussed n the subsectons below. 3. Intal Metrc Step Doman experts are asked to dentfy components (.e., dstance metrcs) applcable to the graphs that wll serve as buldng blocks for the learnng of a new metrc. If the experts know the relatve mportance of the components, ths nformaton s used to assgn ntal weghts. An Intal Weght Heurstc s proposed. Intal Weght Heurstc: Assgn ntal weghts to the components n the LearnMet dstance metrc based on the relatve mportance of the components on the graphs n the doman.

3 If the components relatve mportance s unknown, random weghts are assgned. Intal weghts are on a scale of to 0. : INITIAL METRIC STEP GIVEN: DOMAIN EXPERT INPUT ON DISTANCE TYPES FOR EACH DISTANCE TYPE ASSIGN A COMPONENT TO D IF RELATIVE IMPORTANCE OF COMPONENTS IS AVAILABLE THEN USE INITIAL WEIGHT HEURISTIC ELSE ASSIGN RANDOM WEIGHTS TO COMPONENTS 3. Clusterng Step Usng D as the dstance metrc, k clusters are constructed usng an arbtrary but fxed clusterng algorthm (e.g., k-means [0]), where k s the number of clusters n the tranng set. : CLUSTERING STEP GIVEN: A TRAINING SET CONSISTING OF A COLLECTION OF GRAPHS AND A CORRECT k-clustering OF THEM SELECT AN ARBITRARY BUT FIXED CLUSTERING ALGORITHM SET NUMBER OF CLUSTERS TO k ( CONSTANT) CLUSTER GRAPHS USING D = wdc w m Dc m 3.3 Cluster Evaluaton Step The clusters obtaned from the algorthm,.e., the predcted clusters, are evaluated by comparng them wth correct clusters n the tranng set,.e., the actual clusters. An example of predcted and actual clusters s shown n Fgure. Ideally, the predcted clusters should match the actual clusters. Any dfference between predcted and actual clusters s consdered an error. To compute ths error, we consder pars of graphs and ntroduce the followng notaton. True/False Postve/Negatve Pars of Graphs: Gven a par of graphs I and J, we say that: (I,J) s a True Postve (TP) par f I and J are n the same actual cluster and n the same predcted cluster. (I,J) s a True Negatve (TN) par f I and J are n dfferent actual clusters and n dfferent predcted clusters. (I,J) s a False Postve (FP) par f I and J are n dfferent actual clusters but n the same predcted cluster. (I,J) s a False Negatve (FN) par f I and J are n the same actual cluster but n dfferent predcted clusters. Fgure ncludes examples of each of these knds of pars: (g,g ) s a true postve par; (g,g 3 ) s a true negatve par; (g 3,g 4 ) s a false postve par; and (g 4,g 6 ) s a false negatve par. The error measure of nterest to us s falure rate whch s defned below. Success and s: Let TP, TN, FP and FN denote the number of true postve, true negatve, false postve and false negatve pars respectvely. Also let SR denote the Success Rate and FR = ( SR) denote the, as defned below: TP + TN SR = TP + FP + TN + FN FP + FN FR = TP + FP + TN + FN Fgure : Predcted and Actual Clusters In our doman, false postves and false negatves are equally undesrable. Hence, our defnton of falure rate weghs them equally. Gven a number G of graphs n the tranng set, the total number of pars of graphs s G C = G! (!( G )! ). Thus, for 5 graphs there are 300 pars, for 50 graphs, 5 pars, etc. We defne an epoch n LearnMet as one run of all ts steps. That s, a complete tranng cycle. Overfttng: To avod overfttng n LearnMet, we use an approach analogous to ncremental gradent descent [, 4]. Instead of usng all pars of graphs for evaluaton, a subset of pars s used called ppe or pars per epoch. In each epoch, a dstnct combnaton of pars s used for evaluaton and weght adjustments. Thus there s enough randomzaton n every epoch If ppe = 5, then we have a total of C5 =.95 0 dstnct pars for learnng [, 4]. Thus n each epoch 5 dstnct pars can be used. Ths stll gves a large number of epochs wth dstnct pars for learnng. Ths ncremental approach reduces the tme complexty of the algorthm and helps avod overfttng. Determnng the best ppe value s an optmzaton problem. Also n LearnMet, the random seed s altered n the clusterng algorthm n dfferent epochs as an addtonal method to avod overfttng. Ideally, the error.e., falure rate n an epoch should be zero. However, n practce a doman-specfc error threshold t s used. Error Threshold: A doman-specfc error threshold t s the extent of error allowed per epoch n the doman, where error s measured by falure rate. Dstance between a Par of Graphs: The dstance D(g a,g b ) between a par of graphs g a and g b s the weghted sum of components n the graphs usng metrc D. Thus, D g, g ) = w Dc ( g, g ) w Dc ( g, g ) ( a b a b m m a b Gven ths, consder FN pars, e.g., (g 4,g 5 ) and (g 4,g 6 ). These pars are n the same actual cluster. However they are predcted to be n dfferent clusters. Snce predcted clusters are obtaned wth the metrc D, the (average) dstance D(g a,g b ) for these pars s greater than t should be. Conversely, for FP pars n dfferent actual, same predcted clusters, e.g., (g 3,g 4 ), the (average) dstance D(g a,g b ) s smaller than t should be.

4 Average FN and FP Dstances D_FN and D_FP: FN D _ FN = (/ FN ) j = D _ FP = (/ FP) FP j = D ( g g a, b ) where (g a,g b ) denotes FN pars. D( g a, g b ) where (g a,g b ) denotes FP pars. 3: CLUSTER EVALUATION STEP SET ppe TO THE DESIRED NUMBER OF PAIRS OF GRAPHS FROM THE TRAINING DATASET TO BE CONSIDERED IN AN EPOCH RANDOMLY SELECT ppe PAIRS OF GRAPHS SET THE ERROR THRESHOLD t IDENTIFY THE TP, TN, FP, FN FROM ppe PAIRS CALCULATE FAILURE RATE FR IF (FR < t) THEN RETURN CLUSTERING IS ACCURATE ELSE CALCULATE D_FN,, D_FP 3.4 Weght Adjustment Step If the result of the evaluaton does not ndcate that the clusterng s accurate, then the dstances D_FN and D_FP are used to make weght adjustments to reduce the error n clusterng. Consder the error n FN pars. To reduce the average error D_FN, the weghts of one or more components n the metrc used to calculate the dstance n the present epoch s decreased. For ths we propose the FN Heurstc. FN Heurstc: Decrease the weghts of the components n the metrc D n proporton to ther contrbutons to the dstance D_FN. That s, for each component: D _ FNc w ' = w where D _ FNc = D _ FN for Dc alone D _ FN Conversely, consder FP pars. To reduce ther error, we ncrease D_FP. Ths s done by ncreasng the weghts of one or more components n the metrc usng the FP Heurstc. FP Heurstc: Increase the weghts of the components n the metrc D n proporton to ther contrbutons to the dstance D_FP. That s, for each component: D _ FPc w '' = w + where D _ FPc = D _ FP for Dc alone D _ FP Combnng these two adjustments: Weght Adjustment Heurstc: D _ FNc w ''' = max(0, w ( ) D _ FN + D _ FPc ( )) D _ FP Thus D ''' = w ''' Dc w m ''' Dcm The new metrc D obtaned after weght adjustments s lkely to mnmze the error due to the FN and FP type pars. If the weght of a component becomes negatve, t s converted to zero as we consder only non-negatve weghts. Clusterng s done wth ths new metrc. If the resultng error s below the threshold, then a confrmatory test usng the same metrc to cluster for more epochs s performed, and then ths step s complete. 4: WEIGHT ADJUSTMENT STEP IF CLUSTERING IS ACCURATE OR MAX EPOCHS REACHED THEN GO TO 5: FINAL METRIC STEP ELSE APPLY WEIGHT ADJUSTMENT HEURISTIC TO GET D GO TO : CLUSTERING STEP 3.5 Fnal Metrc Step If the weght adjustment termnates because the error s below the threshold then the metrc n the last epoch s consdered accurate and t s returned as output. However f termnaton occurs because the maxmum number of epochs s reached, then the most reasonable metrc to be output s the one correspondng to the epoch wth the mnmum error among all epochs. 5: FINAL METRIC STEP IF (FR < t) THEN RETURN METRIC D ELSE FIND EPOCH WITH MINIMUM RETURN CORRESPONDING METRIC D Convergence: LearnMet s not guaranteed to converge or to yeld an optmal dstance metrc. However, thorough expermental evaluaton n our applcaton doman has shown consstent convergence to errors below the requred threshold. 4. EXPERIMENTAL EVALUATION 4. Evaluaton of LearnMet. Expermental Parameters: A tranng set of 300 pars of graphs n Heat Treatng s obtaned from correct clusters of 5 graphs gven by experts. A dstnct test set of 300 pars of graphs s derved from 5 graphs gven by experts. We select ppe = whch yelds C 5 =.95 0 total dstnct combnatons of pars. Thus 5 dstnct pars are used n each epoch. Experts gve an error threshold of 0%,.e., 0. for estmaton. We use the same threshold for clusterng. Intal components n the metrc are gven by experts. Two dstnct assgnments of ntal weghts are gven by two dfferent experts [4]. The correspondng two metrcs are denoted by DE and DE respectvely. A thrd ntal metrc EQU s obtaned by assgnng equal weghts to all components. Several experments are run by assgnng random weghts to components n the ntal metrc [4]. We present two experments wth randomly generated metrcs called RND and RND. See Table. Table : Intal Metrcs n LearnMet Experments Table : Learned Metrcs and Number to Learn

5 Observatons durng Tranng: Table shows the metrc learned n each experment wth number of epochs. Fgures 3 to 7 depct the behavor of LearnMet durng tranng. Experments EQU, RND and RND take longer to converge than DE and DE. However they all converge to approxmately the same D Error over Tranng Set Fgure 3: Experment DE Error over Tranng Set Observatons durng Testng: The learned metrc n each experment s used to cluster graphs n the test set. These clusters are compared wth correct clusters over the test set. Eucldean dstance (ED) s also used to cluster the graphs and the clusters are compared wth the correct clusters. The observatons for all the experments are shown n Fgure 8. Ths fgure depcts the accuracy (success rate) for each metrc over the test set. Clusterng accuraces of the learned metrcs are hgher. Metrc ED RND RND EQU DE DE Clusterng Accuracy over Test Set Accuracy (Success Rate) Accuracy Fgure 4: Experment DE Error over Tranng Set Fgure 5: Experment EQU Fgure 8: Test Set Observatons 4. Evaluaton wth System Integraton Purpose of Integraton: LearnMet has been developed manly for the AutoDomanMne system, whch performs computatonal estmaton of process parameters [3]. AutoDomanMne estmates the graphcal result of a scentfc experment gven ts nput condtons, usng exstng data to predct future trends. Ths technque clusters graphs from exstng experments and sends clusterng output to a decson tree classfer e.g., ID3/J4.8 [,5] to learn the clusterng crtera. For each graph, the nput condtons of ts correspondng experment and the cluster n whch t was placed are used to construct the decson tree. The tree dentfes the combnaton of nput condtons that characterze each cluster. A representatve graph s also selected per cluster. When nput condtons of a new experment are submtted, the tree s traversed to predct the cluster. The representatve graph of that cluster s the estmated graph for that experment. Error over Tranng Set Fgure 6: Experment RND Error over Tranng Set Fgure 9a: Evaluaton wth AutoDomanMne Step Fgure 7: Experment RND

6 Fgure 9b: Evaluaton wth AutoDomanMne Step Evaluatng LearnMet wth AutoDomanMne: LearnMet s evaluated by measurng the accuracy of the AutoDomanMne estmaton wth and wthout the learned metrcs. Ths process s llustrated n Fgures 9a and 9b. The estmaton obtaned from clusterng usng the learned metrcs s compared wth that from clusterng usng Eucldean dstance. Another crteron for comparson s D = DEucldean + DLF + D whch s called the BP AutoDomanMne metrc denoted as ADM [3,4]. Observatons wth AutoDomanMne: The average estmaton accuracy over 0 experments, each usng 0-fold cross valdaton, s shown n Fgure 0. Accuracy wth each metrc output from LearnMet s hgher than that wth Eucldean dstance. Accuraces wth the learned metrcs are also hgher than the accuracy wth the AutoDomanMne metrc. 9% 8% 7% 6% 5% 4% 3% % % % Estmaton Accuracy wth Dfferent Metrcs used n Clusterng ED ADM DE DE EQU RND RND Average Estmaton Accuracy Fgure 0: Evaluaton Results wth AutoDomanMne 5. CONCLUSIONS A technque called LearnMet s proposed to learn a domanspecfc dstance metrc for mnng graphs. LearnMet compares clusters of graphs obtaned from a state-of-the-art clusterng algorthm wth correct clusters gven by experts. An ntal metrc s guessed and refned wth every round of clusterng to gve a fnal metrc wth error below a threshold. LearnMet s evaluated rgorously n the Heat Treatng doman that nspred ts development. Addtonal evaluatons wll be conducted n related domans. Ongong research ncludes determnng a good number of pars per epoch; usng normalzed weghts; consderng scalng factors n weght adjustments; refnng thresholds; and assgnng sutable components to the ntal metrc wthout expert nput. ACKNOWLEDGMENTS Ths work s supported by the Center for Heat Treatng Excellence and by DOE ITP Award DE-FC-07-0ID497. REFERENCES [] Anscombe, F.J. Graphs n Statstcal Analyss, n Amercan Statstcan, 7, 7-, 973. [] Bshop, C.M. Neural Networks for Pattern Recognton, Oxford Unversty Press, England, 996. [3] Boyer, H., and Cary, P. Quenchng and Control of Dstorton, ASM Internatonal, OH, USA, 989. [4] Fourer, J. The Analytcal Theory of Heat (translated by A. Freeman) Dover Publcatons Inc., NY, USA, 955. [5] Fredberg, R.M. A Learnng Machne: Part I, n IBM Journal,, -3, 958. [6] Hnneburg, A., Aggarwal, C., and Kem, D. What s the Nearest Neghbor n Hgh Dmensonal Spaces, n Proceedngs of VLDB Caro, Egypt. August 000. [7] Han, J., and Kamber, M. Data Mnng: Concepts and Technques, Morgan Kaufmann Publshers, CA, USA. 00. [8] Kem, D., and Bustos, B. Smlarty Search n Multmeda Databases, n Proceedngs of ICDE Boston, MA, USA, March 004. [9] Kaufman, L., and Rousseeuw, P. Fndng Groups n Data: An Introducton to Cluster Analyss, John Wley, NY, USA [0] MacQueen, J. B. Some Methods for Classfcaton and Analyss of Multvarate Observatons, n Proceedngs of 5-th Berkeley Symposum on Mathematcal Statstcs and Probablty. :8-97. Berkeley, CA, USA, 967. [] Petrucell, J., Nandram, B., and Chen, M., Appled Statstcs for Engneers and Scentsts, Prentce Hall, NJ, USA, 999. [] Qunlan, J. R. Inducton of Decson Trees, n Machne Learnng, : [3] Varde, A., Rundenstener, E., Ruz, C., Manruzzaman, M., and Ssson Jr., R. Data Mnng over Graphcal Results of Experments wth Doman Semantcs, n Proceedngs of ACM SIGART's nd Internatonal Conference on Intellgent Computng and Informaton Systems ICICIS Caro, Egypt, March 005. [4] Varde, A., Rundenstener, E., Ruz, C., Manruzzaman, M., and Ssson, Jr., R. The LearnMet Approach for Learnng a Doman-Specfc Dstance Metrc to preserve the Semantc Content n Mnng Graphcal Data, TR#: CS-CHTE , WPI, Worcester, MA, USA, June 005. [5] Wtten, I., and Frank, E., Data Mnng: Practcal Machne Learnng Algorthms wth Java Implementatons, Morgan Kaufmann Publshers, CA, USA, 000. [6] Xng, E., Ng., A., Jordan, M., and Russell, S. Dstance Metrc Learnng wth Applcaton to Clusterng wth Sde Informaton, n Proceedngs of NIPS-03, Vancouver, Canada, December 003.

Subspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points;

Subspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points; Subspace clusterng Clusterng Fundamental to all clusterng technques s the choce of dstance measure between data ponts; D q ( ) ( ) 2 x x = x x, j k = 1 k jk Squared Eucldean dstance Assumpton: All features

More information

Content Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers

Content Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers IOSR Journal of Electroncs and Communcaton Engneerng (IOSR-JECE) e-issn: 78-834,p- ISSN: 78-8735.Volume 9, Issue, Ver. IV (Mar - Apr. 04), PP 0-07 Content Based Image Retreval Usng -D Dscrete Wavelet wth

More information

Cluster Analysis of Electrical Behavior

Cluster Analysis of Electrical Behavior Journal of Computer and Communcatons, 205, 3, 88-93 Publshed Onlne May 205 n ScRes. http://www.scrp.org/ournal/cc http://dx.do.org/0.4236/cc.205.350 Cluster Analyss of Electrcal Behavor Ln Lu Ln Lu, School

More information

Problem Definitions and Evaluation Criteria for Computational Expensive Optimization

Problem Definitions and Evaluation Criteria for Computational Expensive Optimization Problem efntons and Evaluaton Crtera for Computatonal Expensve Optmzaton B. Lu 1, Q. Chen and Q. Zhang 3, J. J. Lang 4, P. N. Suganthan, B. Y. Qu 6 1 epartment of Computng, Glyndwr Unversty, UK Faclty

More information

Classifier Selection Based on Data Complexity Measures *

Classifier Selection Based on Data Complexity Measures * Classfer Selecton Based on Data Complexty Measures * Edth Hernández-Reyes, J.A. Carrasco-Ochoa, and J.Fco. Martínez-Trndad Natonal Insttute for Astrophyscs, Optcs and Electroncs, Lus Enrque Erro No.1 Sta.

More information

Learning the Kernel Parameters in Kernel Minimum Distance Classifier

Learning the Kernel Parameters in Kernel Minimum Distance Classifier Learnng the Kernel Parameters n Kernel Mnmum Dstance Classfer Daoqang Zhang 1,, Songcan Chen and Zh-Hua Zhou 1* 1 Natonal Laboratory for Novel Software Technology Nanjng Unversty, Nanjng 193, Chna Department

More information

S1 Note. Basis functions.

S1 Note. Basis functions. S1 Note. Bass functons. Contents Types of bass functons...1 The Fourer bass...2 B-splne bass...3 Power and type I error rates wth dfferent numbers of bass functons...4 Table S1. Smulaton results of type

More information

Outline. Type of Machine Learning. Examples of Application. Unsupervised Learning

Outline. Type of Machine Learning. Examples of Application. Unsupervised Learning Outlne Artfcal Intellgence and ts applcatons Lecture 8 Unsupervsed Learnng Professor Danel Yeung danyeung@eee.org Dr. Patrck Chan patrckchan@eee.org South Chna Unversty of Technology, Chna Introducton

More information

The Research of Support Vector Machine in Agricultural Data Classification

The Research of Support Vector Machine in Agricultural Data Classification The Research of Support Vector Machne n Agrcultural Data Classfcaton Le Sh, Qguo Duan, Xnmng Ma, Me Weng College of Informaton and Management Scence, HeNan Agrcultural Unversty, Zhengzhou 45000 Chna Zhengzhou

More information

Feature Reduction and Selection

Feature Reduction and Selection Feature Reducton and Selecton Dr. Shuang LIANG School of Software Engneerng TongJ Unversty Fall, 2012 Today s Topcs Introducton Problems of Dmensonalty Feature Reducton Statstc methods Prncpal Components

More information

MULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION

MULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION MULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION Paulo Quntlano 1 & Antono Santa-Rosa 1 Federal Polce Department, Brasla, Brazl. E-mals: quntlano.pqs@dpf.gov.br and

More information

Unsupervised Learning

Unsupervised Learning Pattern Recognton Lecture 8 Outlne Introducton Unsupervsed Learnng Parametrc VS Non-Parametrc Approach Mxture of Denstes Maxmum-Lkelhood Estmates Clusterng Prof. Danel Yeung School of Computer Scence and

More information

Hybridization of Expectation-Maximization and K-Means Algorithms for Better Clustering Performance

Hybridization of Expectation-Maximization and K-Means Algorithms for Better Clustering Performance BULGARIAN ACADEMY OF SCIENCES CYBERNETICS AND INFORMATION TECHNOLOGIES Volume 16, No 2 Sofa 2016 Prnt ISSN: 1311-9702; Onlne ISSN: 1314-4081 DOI: 10.1515/cat-2016-0017 Hybrdzaton of Expectaton-Maxmzaton

More information

Compiler Design. Spring Register Allocation. Sample Exercises and Solutions. Prof. Pedro C. Diniz

Compiler Design. Spring Register Allocation. Sample Exercises and Solutions. Prof. Pedro C. Diniz Compler Desgn Sprng 2014 Regster Allocaton Sample Exercses and Solutons Prof. Pedro C. Dnz USC / Informaton Scences Insttute 4676 Admralty Way, Sute 1001 Marna del Rey, Calforna 90292 pedro@s.edu Regster

More information

Machine Learning. Topic 6: Clustering

Machine Learning. Topic 6: Clustering Machne Learnng Topc 6: lusterng lusterng Groupng data nto (hopefully useful) sets. Thngs on the left Thngs on the rght Applcatons of lusterng Hypothess Generaton lusters mght suggest natural groups. Hypothess

More information

An Entropy-Based Approach to Integrated Information Needs Assessment

An Entropy-Based Approach to Integrated Information Needs Assessment Dstrbuton Statement A: Approved for publc release; dstrbuton s unlmted. An Entropy-Based Approach to ntegrated nformaton Needs Assessment June 8, 2004 Wllam J. Farrell Lockheed Martn Advanced Technology

More information

Determining the Optimal Bandwidth Based on Multi-criterion Fusion

Determining the Optimal Bandwidth Based on Multi-criterion Fusion Proceedngs of 01 4th Internatonal Conference on Machne Learnng and Computng IPCSIT vol. 5 (01) (01) IACSIT Press, Sngapore Determnng the Optmal Bandwdth Based on Mult-crteron Fuson Ha-L Lang 1+, Xan-Mn

More information

Support Vector Machines

Support Vector Machines /9/207 MIST.6060 Busness Intellgence and Data Mnng What are Support Vector Machnes? Support Vector Machnes Support Vector Machnes (SVMs) are supervsed learnng technques that analyze data and recognze patterns.

More information

12/2/2009. Announcements. Parametric / Non-parametric. Case-Based Reasoning. Nearest-Neighbor on Images. Nearest-Neighbor Classification

12/2/2009. Announcements. Parametric / Non-parametric. Case-Based Reasoning. Nearest-Neighbor on Images. Nearest-Neighbor Classification Introducton to Artfcal Intellgence V22.0472-001 Fall 2009 Lecture 24: Nearest-Neghbors & Support Vector Machnes Rob Fergus Dept of Computer Scence, Courant Insttute, NYU Sldes from Danel Yeung, John DeNero

More information

Unsupervised Learning and Clustering

Unsupervised Learning and Clustering Unsupervsed Learnng and Clusterng Why consder unlabeled samples?. Collectng and labelng large set of samples s costly Gettng recorded speech s free, labelng s tme consumng 2. Classfer could be desgned

More information

Classifying Acoustic Transient Signals Using Artificial Intelligence

Classifying Acoustic Transient Signals Using Artificial Intelligence Classfyng Acoustc Transent Sgnals Usng Artfcal Intellgence Steve Sutton, Unversty of North Carolna At Wlmngton (suttons@charter.net) Greg Huff, Unversty of North Carolna At Wlmngton (jgh7476@uncwl.edu)

More information

A Unified Framework for Semantics and Feature Based Relevance Feedback in Image Retrieval Systems

A Unified Framework for Semantics and Feature Based Relevance Feedback in Image Retrieval Systems A Unfed Framework for Semantcs and Feature Based Relevance Feedback n Image Retreval Systems Ye Lu *, Chunhu Hu 2, Xngquan Zhu 3*, HongJang Zhang 2, Qang Yang * School of Computng Scence Smon Fraser Unversty

More information

NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS

NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS ARPN Journal of Engneerng and Appled Scences 006-017 Asan Research Publshng Network (ARPN). All rghts reserved. NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS Igor Grgoryev, Svetlana

More information

Machine Learning: Algorithms and Applications

Machine Learning: Algorithms and Applications 14/05/1 Machne Learnng: Algorthms and Applcatons Florano Zn Free Unversty of Bozen-Bolzano Faculty of Computer Scence Academc Year 011-01 Lecture 10: 14 May 01 Unsupervsed Learnng cont Sldes courtesy of

More information

Machine Learning 9. week

Machine Learning 9. week Machne Learnng 9. week Mappng Concept Radal Bass Functons (RBF) RBF Networks 1 Mappng It s probably the best scenaro for the classfcaton of two dataset s to separate them lnearly. As you see n the below

More information

Term Weighting Classification System Using the Chi-square Statistic for the Classification Subtask at NTCIR-6 Patent Retrieval Task

Term Weighting Classification System Using the Chi-square Statistic for the Classification Subtask at NTCIR-6 Patent Retrieval Task Proceedngs of NTCIR-6 Workshop Meetng, May 15-18, 2007, Tokyo, Japan Term Weghtng Classfcaton System Usng the Ch-square Statstc for the Classfcaton Subtask at NTCIR-6 Patent Retreval Task Kotaro Hashmoto

More information

Outline. Discriminative classifiers for image recognition. Where in the World? A nearest neighbor recognition example 4/14/2011. CS 376 Lecture 22 1

Outline. Discriminative classifiers for image recognition. Where in the World? A nearest neighbor recognition example 4/14/2011. CS 376 Lecture 22 1 4/14/011 Outlne Dscrmnatve classfers for mage recognton Wednesday, Aprl 13 Krsten Grauman UT-Austn Last tme: wndow-based generc obect detecton basc ppelne face detecton wth boostng as case study Today:

More information

BOOSTING CLASSIFICATION ACCURACY WITH SAMPLES CHOSEN FROM A VALIDATION SET

BOOSTING CLASSIFICATION ACCURACY WITH SAMPLES CHOSEN FROM A VALIDATION SET 1 BOOSTING CLASSIFICATION ACCURACY WITH SAMPLES CHOSEN FROM A VALIDATION SET TZU-CHENG CHUANG School of Electrcal and Computer Engneerng, Purdue Unversty, West Lafayette, Indana 47907 SAUL B. GELFAND School

More information

Smoothing Spline ANOVA for variable screening

Smoothing Spline ANOVA for variable screening Smoothng Splne ANOVA for varable screenng a useful tool for metamodels tranng and mult-objectve optmzaton L. Rcco, E. Rgon, A. Turco Outlne RSM Introducton Possble couplng Test case MOO MOO wth Game Theory

More information

y and the total sum of

y and the total sum of Lnear regresson Testng for non-lnearty In analytcal chemstry, lnear regresson s commonly used n the constructon of calbraton functons requred for analytcal technques such as gas chromatography, atomc absorpton

More information

Learning-Based Top-N Selection Query Evaluation over Relational Databases

Learning-Based Top-N Selection Query Evaluation over Relational Databases Learnng-Based Top-N Selecton Query Evaluaton over Relatonal Databases Lang Zhu *, Wey Meng ** * School of Mathematcs and Computer Scence, Hebe Unversty, Baodng, Hebe 071002, Chna, zhu@mal.hbu.edu.cn **

More information

A Binarization Algorithm specialized on Document Images and Photos

A Binarization Algorithm specialized on Document Images and Photos A Bnarzaton Algorthm specalzed on Document mages and Photos Ergna Kavalleratou Dept. of nformaton and Communcaton Systems Engneerng Unversty of the Aegean kavalleratou@aegean.gr Abstract n ths paper, a

More information

EVALUATION OF THE PERFORMANCES OF ARTIFICIAL BEE COLONY AND INVASIVE WEED OPTIMIZATION ALGORITHMS ON THE MODIFIED BENCHMARK FUNCTIONS

EVALUATION OF THE PERFORMANCES OF ARTIFICIAL BEE COLONY AND INVASIVE WEED OPTIMIZATION ALGORITHMS ON THE MODIFIED BENCHMARK FUNCTIONS Academc Research Internatonal ISS-L: 3-9553, ISS: 3-9944 Vol., o. 3, May 0 EVALUATIO OF THE PERFORMACES OF ARTIFICIAL BEE COLOY AD IVASIVE WEED OPTIMIZATIO ALGORITHMS O THE MODIFIED BECHMARK FUCTIOS Dlay

More information

X- Chart Using ANOM Approach

X- Chart Using ANOM Approach ISSN 1684-8403 Journal of Statstcs Volume 17, 010, pp. 3-3 Abstract X- Chart Usng ANOM Approach Gullapall Chakravarth 1 and Chaluvad Venkateswara Rao Control lmts for ndvdual measurements (X) chart are

More information

Assignment # 2. Farrukh Jabeen Algorithms 510 Assignment #2 Due Date: June 15, 2009.

Assignment # 2. Farrukh Jabeen Algorithms 510 Assignment #2 Due Date: June 15, 2009. Farrukh Jabeen Algorthms 51 Assgnment #2 Due Date: June 15, 29. Assgnment # 2 Chapter 3 Dscrete Fourer Transforms Implement the FFT for the DFT. Descrbed n sectons 3.1 and 3.2. Delverables: 1. Concse descrpton

More information

An Optimal Algorithm for Prufer Codes *

An Optimal Algorithm for Prufer Codes * J. Software Engneerng & Applcatons, 2009, 2: 111-115 do:10.4236/jsea.2009.22016 Publshed Onlne July 2009 (www.scrp.org/journal/jsea) An Optmal Algorthm for Prufer Codes * Xaodong Wang 1, 2, Le Wang 3,

More information

Data Mining: Model Evaluation

Data Mining: Model Evaluation Data Mnng: Model Evaluaton Aprl 16, 2013 1 Issues: Evaluatng Classfcaton Methods Accurac classfer accurac: predctng class label predctor accurac: guessng value of predcted attrbutes Speed tme to construct

More information

TECHNIQUE OF FORMATION HOMOGENEOUS SAMPLE SAME OBJECTS. Muradaliyev A.Z.

TECHNIQUE OF FORMATION HOMOGENEOUS SAMPLE SAME OBJECTS. Muradaliyev A.Z. TECHNIQUE OF FORMATION HOMOGENEOUS SAMPLE SAME OBJECTS Muradalyev AZ Azerbajan Scentfc-Research and Desgn-Prospectng Insttute of Energetc AZ1012, Ave HZardab-94 E-mal:aydn_murad@yahoocom Importance of

More information

Relevance Assignment and Fusion of Multiple Learning Methods Applied to Remote Sensing Image Analysis

Relevance Assignment and Fusion of Multiple Learning Methods Applied to Remote Sensing Image Analysis Assgnment and Fuson of Multple Learnng Methods Appled to Remote Sensng Image Analyss Peter Bajcsy, We-Wen Feng and Praveen Kumar Natonal Center for Supercomputng Applcaton (NCSA), Unversty of Illnos at

More information

Optimizing Document Scoring for Query Retrieval

Optimizing Document Scoring for Query Retrieval Optmzng Document Scorng for Query Retreval Brent Ellwen baellwe@cs.stanford.edu Abstract The goal of ths project was to automate the process of tunng a document query engne. Specfcally, I used machne learnng

More information

SLAM Summer School 2006 Practical 2: SLAM using Monocular Vision

SLAM Summer School 2006 Practical 2: SLAM using Monocular Vision SLAM Summer School 2006 Practcal 2: SLAM usng Monocular Vson Javer Cvera, Unversty of Zaragoza Andrew J. Davson, Imperal College London J.M.M Montel, Unversty of Zaragoza. josemar@unzar.es, jcvera@unzar.es,

More information

A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS

A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS Proceedngs of the Wnter Smulaton Conference M E Kuhl, N M Steger, F B Armstrong, and J A Jones, eds A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS Mark W Brantley Chun-Hung

More information

User Authentication Based On Behavioral Mouse Dynamics Biometrics

User Authentication Based On Behavioral Mouse Dynamics Biometrics User Authentcaton Based On Behavoral Mouse Dynamcs Bometrcs Chee-Hyung Yoon Danel Donghyun Km Department of Computer Scence Department of Computer Scence Stanford Unversty Stanford Unversty Stanford, CA

More information

LearnMet: Learning Domain-Specific Distance Metrics for Plots of Scientific Functions

LearnMet: Learning Domain-Specific Distance Metrics for Plots of Scientific Functions LearnMet: Learning Domain-Specific Distance Metrics for Plots of Scientific Functions Aparna Varde 1,2, Elke Rundensteiner 1, Carolina Ruiz 1, Mohammed Maniruzzaman 2, and Richard Sisson Jr. 2 1 Department

More information

Parallelism for Nested Loops with Non-uniform and Flow Dependences

Parallelism for Nested Loops with Non-uniform and Flow Dependences Parallelsm for Nested Loops wth Non-unform and Flow Dependences Sam-Jn Jeong Dept. of Informaton & Communcaton Engneerng, Cheonan Unversty, 5, Anseo-dong, Cheonan, Chungnam, 330-80, Korea. seong@cheonan.ac.kr

More information

Outline. Self-Organizing Maps (SOM) US Hebbian Learning, Cntd. The learning rule is Hebbian like:

Outline. Self-Organizing Maps (SOM) US Hebbian Learning, Cntd. The learning rule is Hebbian like: Self-Organzng Maps (SOM) Turgay İBRİKÇİ, PhD. Outlne Introducton Structures of SOM SOM Archtecture Neghborhoods SOM Algorthm Examples Summary 1 2 Unsupervsed Hebban Learnng US Hebban Learnng, Cntd 3 A

More information

An Image Fusion Approach Based on Segmentation Region

An Image Fusion Approach Based on Segmentation Region Rong Wang, L-Qun Gao, Shu Yang, Yu-Hua Cha, and Yan-Chun Lu An Image Fuson Approach Based On Segmentaton Regon An Image Fuson Approach Based on Segmentaton Regon Rong Wang, L-Qun Gao, Shu Yang 3, Yu-Hua

More information

A Fast Content-Based Multimedia Retrieval Technique Using Compressed Data

A Fast Content-Based Multimedia Retrieval Technique Using Compressed Data A Fast Content-Based Multmeda Retreval Technque Usng Compressed Data Borko Furht and Pornvt Saksobhavvat NSF Multmeda Laboratory Florda Atlantc Unversty, Boca Raton, Florda 3343 ABSTRACT In ths paper,

More information

Fuzzy Modeling of the Complexity vs. Accuracy Trade-off in a Sequential Two-Stage Multi-Classifier System

Fuzzy Modeling of the Complexity vs. Accuracy Trade-off in a Sequential Two-Stage Multi-Classifier System Fuzzy Modelng of the Complexty vs. Accuracy Trade-off n a Sequental Two-Stage Mult-Classfer System MARK LAST 1 Department of Informaton Systems Engneerng Ben-Guron Unversty of the Negev Beer-Sheva 84105

More information

F Geometric Mean Graphs

F Geometric Mean Graphs Avalable at http://pvamu.edu/aam Appl. Appl. Math. ISSN: 1932-9466 Vol. 10, Issue 2 (December 2015), pp. 937-952 Applcatons and Appled Mathematcs: An Internatonal Journal (AAM) F Geometrc Mean Graphs A.

More information

CS 534: Computer Vision Model Fitting

CS 534: Computer Vision Model Fitting CS 534: Computer Vson Model Fttng Sprng 004 Ahmed Elgammal Dept of Computer Scence CS 534 Model Fttng - 1 Outlnes Model fttng s mportant Least-squares fttng Maxmum lkelhood estmaton MAP estmaton Robust

More information

A Fast Visual Tracking Algorithm Based on Circle Pixels Matching

A Fast Visual Tracking Algorithm Based on Circle Pixels Matching A Fast Vsual Trackng Algorthm Based on Crcle Pxels Matchng Zhqang Hou hou_zhq@sohu.com Chongzhao Han czhan@mal.xjtu.edu.cn Ln Zheng Abstract: A fast vsual trackng algorthm based on crcle pxels matchng

More information

Support Vector Machines

Support Vector Machines Support Vector Machnes Decson surface s a hyperplane (lne n 2D) n feature space (smlar to the Perceptron) Arguably, the most mportant recent dscovery n machne learnng In a nutshell: map the data to a predetermned

More information

FEATURE EXTRACTION. Dr. K.Vijayarekha. Associate Dean School of Electrical and Electronics Engineering SASTRA University, Thanjavur

FEATURE EXTRACTION. Dr. K.Vijayarekha. Associate Dean School of Electrical and Electronics Engineering SASTRA University, Thanjavur FEATURE EXTRACTION Dr. K.Vjayarekha Assocate Dean School of Electrcal and Electroncs Engneerng SASTRA Unversty, Thanjavur613 41 Jont Intatve of IITs and IISc Funded by MHRD Page 1 of 8 Table of Contents

More information

Helsinki University Of Technology, Systems Analysis Laboratory Mat Independent research projects in applied mathematics (3 cr)

Helsinki University Of Technology, Systems Analysis Laboratory Mat Independent research projects in applied mathematics (3 cr) Helsnk Unversty Of Technology, Systems Analyss Laboratory Mat-2.08 Independent research projects n appled mathematcs (3 cr) "! #$&% Antt Laukkanen 506 R ajlaukka@cc.hut.f 2 Introducton...3 2 Multattrbute

More information

SHAPE RECOGNITION METHOD BASED ON THE k-nearest NEIGHBOR RULE

SHAPE RECOGNITION METHOD BASED ON THE k-nearest NEIGHBOR RULE SHAPE RECOGNITION METHOD BASED ON THE k-nearest NEIGHBOR RULE Dorna Purcaru Faculty of Automaton, Computers and Electroncs Unersty of Craoa 13 Al. I. Cuza Street, Craoa RO-1100 ROMANIA E-mal: dpurcaru@electroncs.uc.ro

More information

TN348: Openlab Module - Colocalization

TN348: Openlab Module - Colocalization TN348: Openlab Module - Colocalzaton Topc The Colocalzaton module provdes the faclty to vsualze and quantfy colocalzaton between pars of mages. The Colocalzaton wndow contans a prevew of the two mages

More information

The Greedy Method. Outline and Reading. Change Money Problem. Greedy Algorithms. Applications of the Greedy Strategy. The Greedy Method Technique

The Greedy Method. Outline and Reading. Change Money Problem. Greedy Algorithms. Applications of the Greedy Strategy. The Greedy Method Technique //00 :0 AM Outlne and Readng The Greedy Method The Greedy Method Technque (secton.) Fractonal Knapsack Problem (secton..) Task Schedulng (secton..) Mnmum Spannng Trees (secton.) Change Money Problem Greedy

More information

The Codesign Challenge

The Codesign Challenge ECE 4530 Codesgn Challenge Fall 2007 Hardware/Software Codesgn The Codesgn Challenge Objectves In the codesgn challenge, your task s to accelerate a gven software reference mplementaton as fast as possble.

More information

Meta-heuristics for Multidimensional Knapsack Problems

Meta-heuristics for Multidimensional Knapsack Problems 2012 4th Internatonal Conference on Computer Research and Development IPCSIT vol.39 (2012) (2012) IACSIT Press, Sngapore Meta-heurstcs for Multdmensonal Knapsack Problems Zhbao Man + Computer Scence Department,

More information

Module Management Tool in Software Development Organizations

Module Management Tool in Software Development Organizations Journal of Computer Scence (5): 8-, 7 ISSN 59-66 7 Scence Publcatons Management Tool n Software Development Organzatons Ahmad A. Al-Rababah and Mohammad A. Al-Rababah Faculty of IT, Al-Ahlyyah Amman Unversty,

More information

LECTURE : MANIFOLD LEARNING

LECTURE : MANIFOLD LEARNING LECTURE : MANIFOLD LEARNING Rta Osadchy Some sldes are due to L.Saul, V. C. Raykar, N. Verma Topcs PCA MDS IsoMap LLE EgenMaps Done! Dmensonalty Reducton Data representaton Inputs are real-valued vectors

More information

6.854 Advanced Algorithms Petar Maymounkov Problem Set 11 (November 23, 2005) With: Benjamin Rossman, Oren Weimann, and Pouya Kheradpour

6.854 Advanced Algorithms Petar Maymounkov Problem Set 11 (November 23, 2005) With: Benjamin Rossman, Oren Weimann, and Pouya Kheradpour 6.854 Advanced Algorthms Petar Maymounkov Problem Set 11 (November 23, 2005) Wth: Benjamn Rossman, Oren Wemann, and Pouya Kheradpour Problem 1. We reduce vertex cover to MAX-SAT wth weghts, such that the

More information

Three supervised learning methods on pen digits character recognition dataset

Three supervised learning methods on pen digits character recognition dataset Three supervsed learnng methods on pen dgts character recognton dataset Chrs Flezach Department of Computer Scence and Engneerng Unversty of Calforna, San Dego San Dego, CA 92093 cflezac@cs.ucsd.edu Satoru

More information

1. Introduction. Abstract

1. Introduction. Abstract Image Retreval Usng a Herarchy of Clusters Danela Stan & Ishwar K. Seth Intellgent Informaton Engneerng Laboratory, Department of Computer Scence & Engneerng, Oaland Unversty, Rochester, Mchgan 48309-4478

More information

Intelligent Information Acquisition for Improved Clustering

Intelligent Information Acquisition for Improved Clustering Intellgent Informaton Acquston for Improved Clusterng Duy Vu Unversty of Texas at Austn duyvu@cs.utexas.edu Mkhal Blenko Mcrosoft Research mblenko@mcrosoft.com Prem Melvlle IBM T.J. Watson Research Center

More information

Life Tables (Times) Summary. Sample StatFolio: lifetable times.sgp

Life Tables (Times) Summary. Sample StatFolio: lifetable times.sgp Lfe Tables (Tmes) Summary... 1 Data Input... 2 Analyss Summary... 3 Survval Functon... 5 Log Survval Functon... 6 Cumulatve Hazard Functon... 7 Percentles... 7 Group Comparsons... 8 Summary The Lfe Tables

More information

Learning an Image Manifold for Retrieval

Learning an Image Manifold for Retrieval Learnng an Image Manfold for Retreval Xaofe He*, We-Yng Ma, and Hong-Jang Zhang Mcrosoft Research Asa Bejng, Chna, 100080 {wyma,hjzhang}@mcrosoft.com *Department of Computer Scence, The Unversty of Chcago

More information

BAYESIAN MULTI-SOURCE DOMAIN ADAPTATION

BAYESIAN MULTI-SOURCE DOMAIN ADAPTATION BAYESIAN MULTI-SOURCE DOMAIN ADAPTATION SHI-LIANG SUN, HONG-LEI SHI Department of Computer Scence and Technology, East Chna Normal Unversty 500 Dongchuan Road, Shangha 200241, P. R. Chna E-MAIL: slsun@cs.ecnu.edu.cn,

More information

A Deflected Grid-based Algorithm for Clustering Analysis

A Deflected Grid-based Algorithm for Clustering Analysis A Deflected Grd-based Algorthm for Clusterng Analyss NANCY P. LIN, CHUNG-I CHANG, HAO-EN CHUEH, HUNG-JEN CHEN, WEI-HUA HAO Department of Computer Scence and Informaton Engneerng Tamkang Unversty 5 Yng-chuan

More information

Query Clustering Using a Hybrid Query Similarity Measure

Query Clustering Using a Hybrid Query Similarity Measure Query clusterng usng a hybrd query smlarty measure Fu. L., Goh, D.H., & Foo, S. (2004). WSEAS Transacton on Computers, 3(3), 700-705. Query Clusterng Usng a Hybrd Query Smlarty Measure Ln Fu, Don Hoe-Lan

More information

A mathematical programming approach to the analysis, design and scheduling of offshore oilfields

A mathematical programming approach to the analysis, design and scheduling of offshore oilfields 17 th European Symposum on Computer Aded Process Engneerng ESCAPE17 V. Plesu and P.S. Agach (Edtors) 2007 Elsever B.V. All rghts reserved. 1 A mathematcal programmng approach to the analyss, desgn and

More information

A Multi-step Strategy for Shape Similarity Search In Kamon Image Database

A Multi-step Strategy for Shape Similarity Search In Kamon Image Database A Mult-step Strategy for Shape Smlarty Search In Kamon Image Database Paul W.H. Kwan, Kazuo Torach 2, Kesuke Kameyama 2, Junbn Gao 3, Nobuyuk Otsu 4 School of Mathematcs, Statstcs and Computer Scence,

More information

Simulation: Solving Dynamic Models ABE 5646 Week 11 Chapter 2, Spring 2010

Simulation: Solving Dynamic Models ABE 5646 Week 11 Chapter 2, Spring 2010 Smulaton: Solvng Dynamc Models ABE 5646 Week Chapter 2, Sprng 200 Week Descrpton Readng Materal Mar 5- Mar 9 Evaluatng [Crop] Models Comparng a model wth data - Graphcal, errors - Measures of agreement

More information

NAG Fortran Library Chapter Introduction. G10 Smoothing in Statistics

NAG Fortran Library Chapter Introduction. G10 Smoothing in Statistics Introducton G10 NAG Fortran Lbrary Chapter Introducton G10 Smoothng n Statstcs Contents 1 Scope of the Chapter... 2 2 Background to the Problems... 2 2.1 Smoothng Methods... 2 2.2 Smoothng Splnes and Regresson

More information

Type-2 Fuzzy Non-uniform Rational B-spline Model with Type-2 Fuzzy Data

Type-2 Fuzzy Non-uniform Rational B-spline Model with Type-2 Fuzzy Data Malaysan Journal of Mathematcal Scences 11(S) Aprl : 35 46 (2017) Specal Issue: The 2nd Internatonal Conference and Workshop on Mathematcal Analyss (ICWOMA 2016) MALAYSIAN JOURNAL OF MATHEMATICAL SCIENCES

More information

Skew Angle Estimation and Correction of Hand Written, Textual and Large areas of Non-Textual Document Images: A Novel Approach

Skew Angle Estimation and Correction of Hand Written, Textual and Large areas of Non-Textual Document Images: A Novel Approach Angle Estmaton and Correcton of Hand Wrtten, Textual and Large areas of Non-Textual Document Images: A Novel Approach D.R.Ramesh Babu Pyush M Kumat Mahesh D Dhannawat PES Insttute of Technology Research

More information

Keywords - Wep page classification; bag of words model; topic model; hierarchical classification; Support Vector Machines

Keywords - Wep page classification; bag of words model; topic model; hierarchical classification; Support Vector Machines (IJCSIS) Internatonal Journal of Computer Scence and Informaton Securty, Herarchcal Web Page Classfcaton Based on a Topc Model and Neghborng Pages Integraton Wongkot Srura Phayung Meesad Choochart Haruechayasak

More information

Empirical Distributions of Parameter Estimates. in Binary Logistic Regression Using Bootstrap

Empirical Distributions of Parameter Estimates. in Binary Logistic Regression Using Bootstrap Int. Journal of Math. Analyss, Vol. 8, 4, no. 5, 7-7 HIKARI Ltd, www.m-hkar.com http://dx.do.org/.988/jma.4.494 Emprcal Dstrbutons of Parameter Estmates n Bnary Logstc Regresson Usng Bootstrap Anwar Ftranto*

More information

UB at GeoCLEF Department of Geography Abstract

UB at GeoCLEF Department of Geography   Abstract UB at GeoCLEF 2006 Mguel E. Ruz (1), Stuart Shapro (2), June Abbas (1), Slva B. Southwck (1) and Davd Mark (3) State Unversty of New York at Buffalo (1) Department of Lbrary and Informaton Studes (2) Department

More information

Sensor Selection with Grey Correlation Analysis for Remaining Useful Life Evaluation

Sensor Selection with Grey Correlation Analysis for Remaining Useful Life Evaluation Sensor Selecton wth Grey Correlaton Analyss for Remanng Useful Lfe valuaton Peng Yu, Xu Yong, Lu Datong, Peng Xyuan Automatc est Control Insttute, Harbn Insttute of echnology, Harbn, 5, Chna pengyu@ht.edu.cn

More information

For instance, ; the five basic number-sets are increasingly more n A B & B A A = B (1)

For instance, ; the five basic number-sets are increasingly more n A B & B A A = B (1) Secton 1.2 Subsets and the Boolean operatons on sets If every element of the set A s an element of the set B, we say that A s a subset of B, or that A s contaned n B, or that B contans A, and we wrte A

More information

EXTENDED BIC CRITERION FOR MODEL SELECTION

EXTENDED BIC CRITERION FOR MODEL SELECTION IDIAP RESEARCH REPORT EXTEDED BIC CRITERIO FOR ODEL SELECTIO Itshak Lapdot Andrew orrs IDIAP-RR-0-4 Dalle olle Insttute for Perceptual Artfcal Intellgence P.O.Box 59 artgny Valas Swtzerland phone +4 7

More information

A PATTERN RECOGNITION APPROACH TO IMAGE SEGMENTATION

A PATTERN RECOGNITION APPROACH TO IMAGE SEGMENTATION 1 THE PUBLISHING HOUSE PROCEEDINGS OF THE ROMANIAN ACADEMY, Seres A, OF THE ROMANIAN ACADEMY Volume 4, Number 2/2003, pp.000-000 A PATTERN RECOGNITION APPROACH TO IMAGE SEGMENTATION Tudor BARBU Insttute

More information

CS246: Mining Massive Datasets Jure Leskovec, Stanford University

CS246: Mining Massive Datasets Jure Leskovec, Stanford University CS46: Mnng Massve Datasets Jure Leskovec, Stanford Unversty http://cs46.stanford.edu /19/013 Jure Leskovec, Stanford CS46: Mnng Massve Datasets, http://cs46.stanford.edu Perceptron: y = sgn( x Ho to fnd

More information

Hermite Splines in Lie Groups as Products of Geodesics

Hermite Splines in Lie Groups as Products of Geodesics Hermte Splnes n Le Groups as Products of Geodescs Ethan Eade Updated May 28, 2017 1 Introducton 1.1 Goal Ths document defnes a curve n the Le group G parametrzed by tme and by structural parameters n the

More information

Correlative features for the classification of textural images

Correlative features for the classification of textural images Correlatve features for the classfcaton of textural mages M A Turkova 1 and A V Gadel 1, 1 Samara Natonal Research Unversty, Moskovskoe Shosse 34, Samara, Russa, 443086 Image Processng Systems Insttute

More information

An Accurate Evaluation of Integrals in Convex and Non convex Polygonal Domain by Twelve Node Quadrilateral Finite Element Method

An Accurate Evaluation of Integrals in Convex and Non convex Polygonal Domain by Twelve Node Quadrilateral Finite Element Method Internatonal Journal of Computatonal and Appled Mathematcs. ISSN 89-4966 Volume, Number (07), pp. 33-4 Research Inda Publcatons http://www.rpublcaton.com An Accurate Evaluaton of Integrals n Convex and

More information

Improvement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration

Improvement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration Improvement of Spatal Resoluton Usng BlockMatchng Based Moton Estmaton and Frame Integraton Danya Suga and Takayuk Hamamoto Graduate School of Engneerng, Tokyo Unversty of Scence, 6-3-1, Nuku, Katsuska-ku,

More information

Solving two-person zero-sum game by Matlab

Solving two-person zero-sum game by Matlab Appled Mechancs and Materals Onlne: 2011-02-02 ISSN: 1662-7482, Vols. 50-51, pp 262-265 do:10.4028/www.scentfc.net/amm.50-51.262 2011 Trans Tech Publcatons, Swtzerland Solvng two-person zero-sum game by

More information

Reducing Frame Rate for Object Tracking

Reducing Frame Rate for Object Tracking Reducng Frame Rate for Object Trackng Pavel Korshunov 1 and We Tsang Oo 2 1 Natonal Unversty of Sngapore, Sngapore 11977, pavelkor@comp.nus.edu.sg 2 Natonal Unversty of Sngapore, Sngapore 11977, oowt@comp.nus.edu.sg

More information

Investigating the Performance of Naïve- Bayes Classifiers and K- Nearest Neighbor Classifiers

Investigating the Performance of Naïve- Bayes Classifiers and K- Nearest Neighbor Classifiers Journal of Convergence Informaton Technology Volume 5, Number 2, Aprl 2010 Investgatng the Performance of Naïve- Bayes Classfers and K- Nearest Neghbor Classfers Mohammed J. Islam *, Q. M. Jonathan Wu,

More information

Virtual Machine Migration based on Trust Measurement of Computer Node

Virtual Machine Migration based on Trust Measurement of Computer Node Appled Mechancs and Materals Onlne: 2014-04-04 ISSN: 1662-7482, Vols. 536-537, pp 678-682 do:10.4028/www.scentfc.net/amm.536-537.678 2014 Trans Tech Publcatons, Swtzerland Vrtual Machne Mgraton based on

More information

Efficient Text Classification by Weighted Proximal SVM *

Efficient Text Classification by Weighted Proximal SVM * Effcent ext Classfcaton by Weghted Proxmal SVM * Dong Zhuang 1, Benyu Zhang, Qang Yang 3, Jun Yan 4, Zheng Chen, Yng Chen 1 1 Computer Scence and Engneerng, Bejng Insttute of echnology, Bejng 100081, Chna

More information

SVM-based Learning for Multiple Model Estimation

SVM-based Learning for Multiple Model Estimation SVM-based Learnng for Multple Model Estmaton Vladmr Cherkassky and Yunqan Ma Department of Electrcal and Computer Engneerng Unversty of Mnnesota Mnneapols, MN 55455 {cherkass,myq}@ece.umn.edu Abstract:

More information

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 15

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 15 CS434a/541a: Pattern Recognton Prof. Olga Veksler Lecture 15 Today New Topc: Unsupervsed Learnng Supervsed vs. unsupervsed learnng Unsupervsed learnng Net Tme: parametrc unsupervsed learnng Today: nonparametrc

More information

Using Neural Networks and Support Vector Machines in Data Mining

Using Neural Networks and Support Vector Machines in Data Mining Usng eural etworks and Support Vector Machnes n Data Mnng RICHARD A. WASIOWSKI Computer Scence Department Calforna State Unversty Domnguez Hlls Carson, CA 90747 USA Abstract: - Multvarate data analyss

More information

Backpropagation: In Search of Performance Parameters

Backpropagation: In Search of Performance Parameters Bacpropagaton: In Search of Performance Parameters ANIL KUMAR ENUMULAPALLY, LINGGUO BU, and KHOSROW KAIKHAH, Ph.D. Computer Scence Department Texas State Unversty-San Marcos San Marcos, TX-78666 USA ae049@txstate.edu,

More information

A Robust Method for Estimating the Fundamental Matrix

A Robust Method for Estimating the Fundamental Matrix Proc. VIIth Dgtal Image Computng: Technques and Applcatons, Sun C., Talbot H., Ourseln S. and Adraansen T. (Eds.), 0- Dec. 003, Sydney A Robust Method for Estmatng the Fundamental Matrx C.L. Feng and Y.S.

More information

Performance Evaluation of Information Retrieval Systems

Performance Evaluation of Information Retrieval Systems Why System Evaluaton? Performance Evaluaton of Informaton Retreval Systems Many sldes n ths secton are adapted from Prof. Joydeep Ghosh (UT ECE) who n turn adapted them from Prof. Dk Lee (Unv. of Scence

More information