Applying Continuous Action Reinforcement Learning Automata(CARLA) to Global Training of Hidden Markov Models
|
|
- Megan Whitehead
- 6 years ago
- Views:
Transcription
1 Applyng Contnuous Acton Renforcement Learnng Automata(CARLA to Global Tranng of Hdden Markov Models Jahanshah Kabudan, Mohammad Reza Meybod, and Mohammad Mehd Homayounpour Department of Computer Engneerng and Informaton Technology, AmrKabr Unversty of Technology (Tehran PolyTechnc, Tehran, Iran {kabudan, meybod, Abstract In ths research, we have employed global search and global optzaton technques based on Sulated Annealng (SA and Contnuous Acton Renforcement Learnng Automata (CARLA for global tranng of Hdden Markov Models. The man goal of ths paper s comparng CARLA method to other contnuous global optzaton methods lke SA. Experental results show that the CARLA outperforms SA. Ths s due to the fact that CARLA s a contnuous global optzaton method wth memory and SA s a memoryless one.. Introducton HMM s one of the most powerful methods for processng and modelng stochastc processes and sequences. The most popular method for tranng HMM s Baum-Welch (BW method that s a local search method and lke the other local methods suffers from local opta problem. To cope wth ths problem, global search and optzaton methods are used for tranng HMM whch can easly escape from local opta. The man goal of ths paper s a comparson of CARLA and SA as two global search methods for tranng HMMs. In secton two, we ntroduce HMM. In secton three, BW tranng method s brefly explaned. In secton four, SA-based and CARLA-based global search method s descrbed. Secton fve represents the results of the conducted experents, and fnally, secton sx concludes ths paper.. Hdden Markov Model HMM s a fnte-state system wth a number of states (. Ths system changes ts states probablstcally from a state to another state accordng to state transton probabltes. HMM has two types[]: Contnuous Densty HMM and Dscrete Densty HMM. Dscrete Densty HMM s rarely used for modelng and processng contnuous observatons. Assume that an observaton sequence wth te length T s to be generated by ths model. Observaton sequence s a sequence of observaton vectors as follows: O { o o... o t... o T } Each of the observaton vectors o t s a D-densonal vector and t s te ndex. Startng the te, for t, HMM goes nto a state (e.g. wth correspondng ntal state probablty( ( s are elements of ntal state probabltes vector. In state, vector o t s generated usng a multdensonal PDF b (o t. If the model s n state at te t-, then t enters nto state j at te t wth probablty a j (a j s are elements of state transton probabltes matrx A. An observaton sequence O s generated by HMM wth probablty P(O. Ths probablty s computed usng efcent methods lke Forward or Backward Procedure[]. In contnuous densty HMM, multdensonal PDF n state,.e. b (o, s usually a mxture of Gaussan (ormal multdensonal PDF lke: M b ( o c ( o, µ, C ( o, µ, C s a D-densonal Gaussan (ormal PDF wth Mean Vector and Covarance Matrx C of ths form: T ( o, µ, C.exp( ( o µ C ( o µ D (π.det( C Each of the observaton vectors o t s a D column vector. It s necessary that volume (hypervolume under ths mxture PDF be equal to, therefore: M c M s number of Gaussan PDFs n a mxture PDF of each state. Also, the followng constrants must be satsfed: π and a j j C - s nverse of Matrx C and det(c s the absolute value of the determnant of Matrx C. HMM model wth parameters, A and b( s shown as follows: λ ( π, A,b( b( n ths trplet s a set of parameters ncludng weghts (c, Mean vectors ( and Covarance matrces (C of Gaussan PDFs. 3. Baum-Welch Algorthm In real-world applcatons of HMM, we have usually one or more tranng observaton sequences generated by an unknown model. The goal s estatng Proceedngs of the Internatonal Conference on Informaton Technology: Codng and Computng (ITCC /04 $ IEEE
2 parameters of HMM model,.e., A and b(. The most popular method for tranng HMMs s BW algorthm. The goal of BW algorthm s fndng a model * such that the probablty of generatng observaton sequence O gven s maxzed[]: λ arg Max P( O λ λ As we know, the BW method s trapped n local maxa. We use global optzaton and search methods to escape from these maxa and fnd the global one of the HMM tranng problem. 4. Global Optzaton and Search Methods To solve hard optzaton and search problems n a near-optal way, global search and optzaton methods are used. Sulated Annealng (SA[], Genetc Algorthms (GA, Evolutonary Strateges (ES and Evolutonary Programmng (EP[3] are of ths category. A wde varety of global optzaton methods exsts[4]. Recently, a new global optzaton method called Contnuous Acton Renforcement Learnng Automaton (CARLA has been proposed for optzaton n contnuous domans[5,6]. Some of these methods (lke SA have a fner resoluton compared to some other methods (lke GA. From the global search methods, we chose SA for comparng to CARLA. 4.. Sulated Annealng Sulated Annealng s a global search (optzaton method whch s based on the laws of thermodynamcs. If the method follows the recommended Temperature Schedule for reducng temperature, then t guarantees fndng the global optum from a mathematcal pont of vew. Ths method has been appled n many applcatons for both contnuous and dscrete problems. SA follows the followng procedure for optzaton: A random ntal pont and a hgh ntal temperature s selected. In each teraton, t generates a new pont around the prevous pont usng a PDF g(x called Generatng Functon. It calculates cost functon E(x for ths pont, and probablstcally accepts ths pont wth probablty h(x whch s called Acceptance Functon. If the new pont s better than the prevous one, the probablty of acceptance wll be more than the probablty of rejecton. Of course, n the begnnng of the search, temperature s hgh and the search regon s wde, and the acceptance and rejecton probabltes are almost equal (0.5. When temperature decreases and the algorthm approaches optum ponts, search wdth s reduced, and accuracy and resoluton of the search are ncreased. In the followng secton, we explan the standard verson of SA,.e. Boltzmann Annealng (BA Boltzmann Annealng Boltzmann Annealng algorthm as follows:. A hgh ntal temperature(t and a random ntal pont (x s selected.. k, x * x 3. A random new pont x s generated around x * usng generatng functon g(x,x * D ( x x g( x, x π. T k.exp Tk 4. Cost functon for pont x,.e. E(x s calculated. 5. ew pont x s accepted wth probablty h(x. (If t s accepted then x * x, else x * doesn t change E E( x E( x h( x E + exp T k 6. kk+ 7. Temperature s reduced accordng to ths temperature schedule T T k. ln k 8. If the stop condton s met then algorthm stops, otherwse t goes to step 3. In the above algorthm, D s number of parameters of the cost functon E(x and x x s the Eucldean dstance between two D-densonal vectors x and x *. Acceptance functon h(x s usually a Barker functon (lke the form mentoned n step 5 of the above algorthm, but Metropols functon can also be used. In standard SA algorthm, generatng functon s of a Gaussan type and the correspondng SA s called Boltzmann Annealng (BA. If temperature schedule (step 7 of algorthm s not faster than nverselogarthmc schedule, then the method statstcally guarantees that t fnds the global optum n an nfnte number of teratons. The BA algorthm wth correspondng temperature schedule s very slow and needs a very hgh number of teratons for approachng global optum. If we use SA for global optzaton and global tranng of HMMs, the cost functon can be defned as follows: E Log P( Oλ 4.. CARLA One of the methods whch s used n optzaton and control, s Learnng Automaton (LA[7]. LA s a model whch nteracts wth an envronment. It apples an acton to the envronment and envronment evaluates ths acton. Accordng to the nputs that t receves from the envronment, LA corrects ts acton selecton mechansm wth a renforcement sgnal, and agan apples new acton to the envronment. Ths process s repeated nfntely or untl a predefned goal s obtaned. In standard models of LA, acton s a dscrete varable. Recently, a Contnuous Acton Renforcement Learnng Automaton (CARLA wth contnuous acton has been presented[5,6]. CARLA can be employed n the envronments wth contnuous parameters for global Proceedngs of the Internatonal Conference on Informaton Technology: Codng and Computng (ITCC /04 $ IEEE
3 optzaton tasks. It s a model for renforcement learnng n contnuous-parameter envronments. CARLA generates new value of acton usng an acton generaton functon called acton generaton PDF. If the number of system (envronment parameters s more than one, e.g. n mnzaton of a multdensonal cost functon, then a populaton of CARLA automata nteract wth the system n parallel and each automaton undertakes optzaton of one of the parameters of the system. These automata haven t any relatonshp or nteracton wth each other. They nteract only wth the system (envronment, and ther ndrect nteracton wth each other s through the system. Each automaton generates a new acton (new value of the parameter n the correspondng denson usng ts acton generaton PDF. System (Envronment, evaluates the set of actons of automata and sends renforcement sgnals to automata. Usng these renforcement sgnals, each CARLA automaton corrects ts acton generaton PDF. Ths process s repeated untl a predefned crteron s acheved. One of the advantages of ths method s that the CARLA has memory and saves many of the prevous ponts n ts memory wth respect to ther relatve portance. CARLA method operates n ths way:. At frst, acton generaton PDFs are ntalzed wth unform dstrbuton PDFs n the range of system parameters.. Usng acton generaton PDFs, each CARLA generates a new value of acton (parameter r. 3. Envronment response s calculated (cost functon n optzaton tasks. 4. A renforcement sgnal s calculated wth respect to the goodness of the envronment response or cost functon. 5. Acton generaton PDFs are corrected usng these renforcement sgnals. 6. If stop crteron s not met, t goes to step, otherwse the algorthm s stopped. Assume varable x s n the range [x mn, x max ]. At frst (n the frst teraton, acton generaton PDF s ntalzed wth ths unform PDF: f ( x, xmax In n-th teraton, acton r s generated usng acton generaton PDF f(x,n: r F( r, n f ( x, n dx z( n For generatng an acton, frstly a unform random number z(n between [0,] s generated. Then, the acton r s selected such that cumulatve dstrbuton functon F(r,n be equal to z(n. Cumulatve dstrbuton functon F(r,n s computed by numercal ntegraton n the nterval [x mn, r]. Samples of f(x,n s stored n memory and t s descretzed. Wherever the value of f(x,n s not avalable, ts value s calculated usng lnear nterpolaton. After generatng new value of acton,.e. r, ths acton s appled to the envronment and envronment response J(n(cost functon n optzaton tasks s obtaned. A renforcement sgnal s calculated for correctng acton generaton PDF: J med J( n β ( n mn max 0,, J med J mn Renforcement sgnal must be n the range [0,]. J med and J mn are Medan and Mnum of costs of R prevously vsted ponts (e.g. R500. If acton generaton PDF n n-th teraton be f(x,n, then the acton generaton PDF n (n+ -th teraton s corrected n ths way: α [ f ( x, n + β ( n H ( x, r] f x [, xmax ] f ( x, n + 0 else H(x,r s a gaussan functon wth mean r of ths form: g h ( x r H ( x, r.exp. ( x ( max g w ( xmax Parameters g h and g w determne the speed and the relatve resoluton (accuracy of the learnng or the search process. We set these parameters to 0.3 and 0.0 respectvely[5,6]. In ths formulaton, s a normalzaton coeffcent whch must be appled to the corrected PDF such that the hypervolume under new acton generaton PDF,.e. f(x,n+ be equal to one. 5. Experents In ths secton, we present the results of experents. Our goal s tranng an HMM wth Maxum Lkelhood (ML crteron by global methods. The HMM to be traned s a contnuous densty HMM wth three states and three Gaussan PDFs per state wth dagonal covarance matrces and two-densonal observaton vectors. Therefore the number of parameters wll be: P( ++M(D+ Wth 3, M3 and D, the number of HMM parameters s 57. Tranng s performed n multple observaton sequences case[], and K0 observaton sequences wth te length T0 are avalable. We want to maxze P(O. O s the set of K observaton sequences for tranng, and s the set of parameters of the HMM model. As we know, HMM tranng s a constraned optzaton problem wth these constrants: v c > 0, a j j, for all Where v s dagonal element of covarance matrx correspondng to -th denson. For convertng ths Proceedngs of the Internatonal Conference on Informaton Technology: Codng and Computng (ITCC /04 $ IEEE
4 constraned optzaton problem to an unconstraned one, we used ths heurstc mappngs: exp( a j aj exp( a c j M j exp( c exp( c v exp( v That s, a method lke SA, optzes the parameters π, a, c, v n an unconstraned manner, but the j converted and the mapped versons of these parameters,.e., a j, c and v, satsfy the constrants of the HMM tranng problem. ow, we explan the experents. 5.. SA In ths secton, SA method s used for global tranng of Hdden Markov Models. Sutable ntal temperatures n dfferent densons were set. Fgure shows a sample performance for BA method. Snce n the BA method, the rate of temperature reducton s very slow even n 0,000 teratons, we cannot reach fne resoluton. Therefore, search area remans wde and the change n the cost functon wll be sudden and consderable even n the end of the search process. Fgure. Sample performance of SA (Log Probablty 5.. CARLA In ths method, Max and Mn of HMM parameters were set wth respect to the tranng data. Each of CARLA automata tres to correct and adapt ts acton generaton PDF, and fnally ts deal acton generaton PDF s a Gaussan-lke PDF centered on the optum value of parameter n that denson. Fgure shows the shape of the acton generaton PDF after 000, 000, 3000 and 0,000 teratons. Fgure. The shape of the acton generaton PDF for the frst parameter of an HMM wth 57 parameters after 000, 000, 3000 and 0000 teratons CARLA has a good performance n optzaton tasks wth a few number of parameters (e.g. or 3, and wth a few number of local opta, but the CARLA performance s severely degraded when the number of parameters or the number of local opta grows, and the CARLA becomes slow. Fgure 3 shows sample performance of CARLA for tranng an HMM wth 57 parameters A Comparson between CARLA and SA Our goal n ths secton s comparng CARLA wth SA method for tranng HMMs. In all the experents, number of teratons s 0,000. For generalzaton and Proceedngs of the Internatonal Conference on Informaton Technology: Codng and Computng (ITCC /04 $ IEEE
5 consstency of the results, each method must be performed many tes and experental condtons must be the same for two methods. For ths purpose, we performed each experent 50 tes. That s, 50 dfferent and random ntal ponts and 50 random dfferent tranng set were generated. Of course, ntal condtons and tranng sets were exactly the same for two methods. Also, we used the same seed for startng random number generaton routnes for two methods. The average performance of each method was consdered as the Mean of Log-Probablty n dfferent executons of that method. Table shows average performances for two methods. [] Krkpatrck, S., Gelatt, C.D., Vecch, M.P., Optzaton by Sulated Annealng, Scence, Vol. 0, o. 4598, pp , May 983. [3] Bäck, T., Evolutonary Algorthms n Theory and Practce, Oxford Unversty Press, 996. [4] Stuckman, B.E., Easom, E.E., A Comparson of Bayesan/Samplng Global Optzaton Technques, IEEE Trans. on SMC, Vol., o. 5, Sep. 99. [5] Howell, M.., Gordon, T.J., Best, M.C., The Applcaton of Contnuous Acton Renforcement Learnng Automata to Adaptve PID Tunng, Proceedngs of IEE Control Semnar on Learnng Systems for Control, pp. -0, Brmngham, UK, May 000. [6] Howell, M.., Gordon, T.J., Contnuous Learnng Automata and Adaptve Dgtal Flter Desgn, UKACC Internatonal Conference on Control, Sep [7] arendra, K.S., Thathatchar, M.A.L., Learnng Automata: An Introducton, Prentce-Hall, 989. Fgure 3. Sample performance of CARLA method for tranng an HMM wth 57 parameters Table. Average Performance of CARLA and SA for 50 dfferent experents CARLA BA As the table shows, CARLA outperforms SA, and ths can be due to the fact that, SA hasn t any memory and has a very slow-decreasng temperature schedule, and the fnal resoluton of SA s much lower than the CARLA resoluton n a moderate number of teratons. 6. Conclusons In ths paper, we used two global optzaton and search technques for tranng hdden Markov models. Among the global optzaton methods, SA method and Contnuous Acton Renforcement Learnng Automaton (CARLA method were evaluated. Experental results show that the CARLA has a hgher performance compared to SA. Ths s due to the fact that CARLA s a contnuous global optzaton method wth memory and SA s a memoryless one. References [] Rabner, L.R., Juang, B.-H., Fundamentals of Speech Recognton, Prentce-Hall, 993. Proceedngs of the Internatonal Conference on Informaton Technology: Codng and Computng (ITCC /04 $ IEEE
An Application of the Dulmage-Mendelsohn Decomposition to Sparse Null Space Bases of Full Row Rank Matrices
Internatonal Mathematcal Forum, Vol 7, 2012, no 52, 2549-2554 An Applcaton of the Dulmage-Mendelsohn Decomposton to Sparse Null Space Bases of Full Row Rank Matrces Mostafa Khorramzadeh Department of Mathematcal
More informationParallelism 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 informationDetermining 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 informationClassifier 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 informationAn 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 informationContent 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 informationLearning 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 informationSubspace 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 informationNUMERICAL 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 informationOptimization Methods: Integer Programming Integer Linear Programming 1. Module 7 Lecture Notes 1. Integer Linear Programming
Optzaton Methods: Integer Prograng Integer Lnear Prograng Module Lecture Notes Integer Lnear Prograng Introducton In all the prevous lectures n lnear prograng dscussed so far, the desgn varables consdered
More informationCluster 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 informationAn Iterative Solution Approach to Process Plant Layout using Mixed Integer Optimisation
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 An Iteratve Soluton Approach to Process Plant Layout usng Mxed
More informationSimulation Based Analysis of FAST TCP using OMNET++
Smulaton Based Analyss of FAST TCP usng OMNET++ Umar ul Hassan 04030038@lums.edu.pk Md Term Report CS678 Topcs n Internet Research Sprng, 2006 Introducton Internet traffc s doublng roughly every 3 months
More informationClassifying 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 informationOptimization of integrated circuits by means of simulated annealing. Jernej Olenšek, Janez Puhan, Árpád Bűrmen, Sašo Tomažič, Tadej Tuma
Optmzaton of ntegrated crcuts by means of smulated annealng Jernej Olenšek, Janez Puhan, Árpád Bűrmen, Sašo Tomažč, Tadej Tuma Unversty of Ljubljana, Faculty of Electrcal Engneerng, Tržaška 25, Ljubljana,
More informationEVALUATION 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 informationUnsupervised 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 informationExercises (Part 4) Introduction to R UCLA/CCPR. John Fox, February 2005
Exercses (Part 4) Introducton to R UCLA/CCPR John Fox, February 2005 1. A challengng problem: Iterated weghted least squares (IWLS) s a standard method of fttng generalzed lnear models to data. As descrbed
More information6.854 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 informationSolving 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 informationSmoothing 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 informationA New Approach For the Ranking of Fuzzy Sets With Different Heights
New pproach For the ankng of Fuzzy Sets Wth Dfferent Heghts Pushpnder Sngh School of Mathematcs Computer pplcatons Thapar Unversty, Patala-7 00 Inda pushpndersnl@gmalcom STCT ankng of fuzzy sets plays
More informationImprovement 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 informationHybridization 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 informationNAG 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 informationSum of Linear and Fractional Multiobjective Programming Problem under Fuzzy Rules Constraints
Australan Journal of Basc and Appled Scences, 2(4): 1204-1208, 2008 ISSN 1991-8178 Sum of Lnear and Fractonal Multobjectve Programmng Problem under Fuzzy Rules Constrants 1 2 Sanjay Jan and Kalash Lachhwan
More informationOutline. 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 informationLecture 5: Probability Distributions. Random Variables
Lecture 5: Probablty Dstrbutons Random Varables Probablty Dstrbutons Dscrete Random Varables Contnuous Random Varables and ther Dstrbutons Dscrete Jont Dstrbutons Contnuous Jont Dstrbutons Independent
More informationA 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 informationParameter estimation for incomplete bivariate longitudinal data in clinical trials
Parameter estmaton for ncomplete bvarate longtudnal data n clncal trals Naum M. Khutoryansky Novo Nordsk Pharmaceutcals, Inc., Prnceton, NJ ABSTRACT Bvarate models are useful when analyzng longtudnal data
More informationUser 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 informationDevelopment of Face Tracking and Recognition Algorithm for DVR (Digital Video Recorder)
IJCSNS Internatonal Journal of Computer Scence and Network Securty, VOL.6 No.3A, March 2006 7 Development of Face Trackng and Recognton Algorthm for DVR (Dgtal Vdeo Recorder) Jang-Seon Ryu and Eung-Tae
More informationBAYESIAN 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 informationLoad Balancing for Hex-Cell Interconnection Network
Int. J. Communcatons, Network and System Scences,,, - Publshed Onlne Aprl n ScRes. http://www.scrp.org/journal/jcns http://dx.do.org/./jcns.. Load Balancng for Hex-Cell Interconnecton Network Saher Manaseer,
More informationType-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 informationThe 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 informationPositive Semi-definite Programming Localization in Wireless Sensor Networks
Postve Sem-defnte Programmng Localzaton n Wreless Sensor etworks Shengdong Xe 1,, Jn Wang, Aqun Hu 1, Yunl Gu, Jang Xu, 1 School of Informaton Scence and Engneerng, Southeast Unversty, 10096, anjng Computer
More informationUnsupervised 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 informationS1 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 informationSLAM 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 informationA new paradigm of fuzzy control point in space curve
MATEMATIKA, 2016, Volume 32, Number 2, 153 159 c Penerbt UTM Press All rghts reserved A new paradgm of fuzzy control pont n space curve 1 Abd Fatah Wahab, 2 Mohd Sallehuddn Husan and 3 Mohammad Izat Emr
More informationOPTIMIZATION OF FUZZY RULE BASES USING CONTINUOUS ANT COLONY SYSTEM
Proceedng of the Frst Internatonal Conference on Modelng, Smulaton and Appled Optmzaton, Sharah, U.A.E. February -3, 005 OPTIMIZATION OF FUZZY RULE BASES USING CONTINUOUS ANT COLONY SYSTEM Had Nobahar
More informationInvariant Shape Object Recognition Using B-Spline, Cardinal Spline, and Genetic Algorithm
Proceedngs of the 5th WSEAS Int. Conf. on Sgnal Processng, Robotcs and Automaton, Madrd, Span, February 5-7, 6 (pp4-45) Invarant Shape Obect Recognton Usng B-Splne, Cardnal Splne, and Genetc Algorthm PISIT
More information2x x l. Module 3: Element Properties Lecture 4: Lagrange and Serendipity Elements
Module 3: Element Propertes Lecture : Lagrange and Serendpty Elements 5 In last lecture note, the nterpolaton functons are derved on the bass of assumed polynomal from Pascal s trangle for the fled varable.
More informationA Learning Automata based Algorithm for Solving Traveling Salesman Problem improved by Frequency-based Pruning
Internatonal Journal of Computer Applcatons (0975 8887) Volume 46 No.7, May 202 A Learnng Automata based Algorthm for Solvng Travelng Salesman Problem mproved by Frequencybased Prunng Mr Mohammad Alpour
More informationComparison of Heuristics for Scheduling Independent Tasks on Heterogeneous Distributed Environments
Comparson of Heurstcs for Schedulng Independent Tasks on Heterogeneous Dstrbuted Envronments Hesam Izakan¹, Ath Abraham², Senor Member, IEEE, Václav Snášel³ ¹ Islamc Azad Unversty, Ramsar Branch, Ramsar,
More informationA 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 informationCracking of the Merkle Hellman Cryptosystem Using Genetic Algorithm
Crackng of the Merkle Hellman Cryptosystem Usng Genetc Algorthm Zurab Kochladze 1 * & Lal Besela 2 1 Ivane Javakhshvl Tbls State Unversty, 1, I.Chavchavadze av 1, 0128, Tbls, Georga 2 Sokhum State Unversty,
More informationFace Recognition University at Buffalo CSE666 Lecture Slides Resources:
Face Recognton Unversty at Buffalo CSE666 Lecture Sldes Resources: http://www.face-rec.org/algorthms/ Overvew of face recognton algorthms Correlaton - Pxel based correspondence between two face mages Structural
More informationAn 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 informationEECS 730 Introduction to Bioinformatics Sequence Alignment. Luke Huan Electrical Engineering and Computer Science
EECS 730 Introducton to Bonformatcs Sequence Algnment Luke Huan Electrcal Engneerng and Computer Scence http://people.eecs.ku.edu/~huan/ HMM Π s a set of states Transton Probabltes a kl Pr( l 1 k Probablty
More informationTsinghua University at TAC 2009: Summarizing Multi-documents by Information Distance
Tsnghua Unversty at TAC 2009: Summarzng Mult-documents by Informaton Dstance Chong Long, Mnle Huang, Xaoyan Zhu State Key Laboratory of Intellgent Technology and Systems, Tsnghua Natonal Laboratory for
More informationAn Algorithm for Weighted Positive Influence Dominating Set Based on Learning Automata
4 th Internatonal Conference on Knowledge-Based Engneerng and Innovaton (KBEI-2017) Dec. 22 th, 2017 (Iran Unversty of Scence and Technology) Tehran, Iran An Algorthm for Weghted Postve Influence Domnatng
More informationImproving Low Density Parity Check Codes Over the Erasure Channel. The Nelder Mead Downhill Simplex Method. Scott Stransky
Improvng Low Densty Party Check Codes Over the Erasure Channel The Nelder Mead Downhll Smplex Method Scott Stransky Programmng n conjuncton wth: Bors Cukalovc 18.413 Fnal Project Sprng 2004 Page 1 Abstract
More informationNon-Split Restrained Dominating Set of an Interval Graph Using an Algorithm
Internatonal Journal of Advancements n Research & Technology, Volume, Issue, July- ISS - on-splt Restraned Domnatng Set of an Interval Graph Usng an Algorthm ABSTRACT Dr.A.Sudhakaraah *, E. Gnana Deepka,
More informationA 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 informationLearning-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 informationR s s f. m y s. SPH3UW Unit 7.3 Spherical Concave Mirrors Page 1 of 12. Notes
SPH3UW Unt 7.3 Sphercal Concave Mrrors Page 1 of 1 Notes Physcs Tool box Concave Mrror If the reflectng surface takes place on the nner surface of the sphercal shape so that the centre of the mrror bulges
More informationMeta-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 informationEdge Detection in Noisy Images Using the Support Vector Machines
Edge Detecton n Nosy Images Usng the Support Vector Machnes Hlaro Gómez-Moreno, Saturnno Maldonado-Bascón, Francsco López-Ferreras Sgnal Theory and Communcatons Department. Unversty of Alcalá Crta. Madrd-Barcelona
More informationComplex Filtering and Integration via Sampling
Overvew Complex Flterng and Integraton va Samplng Sgnal processng Sample then flter (remove alases) then resample onunform samplng: jtterng and Posson dsk Statstcs Monte Carlo ntegraton and probablty theory
More informationDetection of an Object by using Principal Component Analysis
Detecton of an Object by usng Prncpal Component Analyss 1. G. Nagaven, 2. Dr. T. Sreenvasulu Reddy 1. M.Tech, Department of EEE, SVUCE, Trupath, Inda. 2. Assoc. Professor, Department of ECE, SVUCE, Trupath,
More informationMULTISPECTRAL 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 informationISSN: International Journal of Engineering and Innovative Technology (IJEIT) Volume 1, Issue 4, April 2012
Performance Evoluton of Dfferent Codng Methods wth β - densty Decodng Usng Error Correctng Output Code Based on Multclass Classfcaton Devangn Dave, M. Samvatsar, P. K. Bhanoda Abstract A common way to
More informationSome Advanced SPC Tools 1. Cumulative Sum Control (Cusum) Chart For the data shown in Table 9-1, the x chart can be generated.
Some Advanced SP Tools 1. umulatve Sum ontrol (usum) hart For the data shown n Table 9-1, the x chart can be generated. However, the shft taken place at sample #21 s not apparent. 92 For ths set samples,
More informationIrregular Cellular Learning Automata and Its Application to Clustering in Sensor Networks
Irregular Cellular Learnng Automata and Its Applcaton to Clusterng n Sensor Networks M. Esnaashar 1, M. R. Meybod 1,2 1 Soft Computng Laboratory, Computer Engneerng and Informaton Technology Department
More informationSupport 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 informationA Study on Clustering for Clustering Based Image De-Noising
Journal of Informaton Systems and Telecommuncaton, Vol. 2, No. 4, October-December 2014 196 A Study on Clusterng for Clusterng Based Image De-Nosng Hossen Bakhsh Golestan* Department of Electrcal Engneerng,
More informationAn 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 informationA 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 informationA 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 informationParallel Numerics. 1 Preconditioning & Iterative Solvers (From 2016)
Technsche Unverstät München WSe 6/7 Insttut für Informatk Prof. Dr. Thomas Huckle Dpl.-Math. Benjamn Uekermann Parallel Numercs Exercse : Prevous Exam Questons Precondtonng & Iteratve Solvers (From 6)
More informationA Simple and Efficient Goal Programming Model for Computing of Fuzzy Linear Regression Parameters with Considering Outliers
62626262621 Journal of Uncertan Systems Vol.5, No.1, pp.62-71, 211 Onlne at: www.us.org.u A Smple and Effcent Goal Programmng Model for Computng of Fuzzy Lnear Regresson Parameters wth Consderng Outlers
More informationSeeking multi-thresholds for image segmentation with Learning Automata
lease cte ths artcle as: Cuevas, E., Zaldvar, D., érez-csneros, M. Seeng mult-thresholds for mage segmentaton wth Learnng Automata, Machne Vson and Applcatons (5), (0), pp. 805-88. Seeng mult-thresholds
More informationIncremental Learning with Support Vector Machines and Fuzzy Set Theory
The 25th Workshop on Combnatoral Mathematcs and Computaton Theory Incremental Learnng wth Support Vector Machnes and Fuzzy Set Theory Yu-Mng Chuang 1 and Cha-Hwa Ln 2* 1 Department of Computer Scence and
More informationFEATURE 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 informationAvailable online at Available online at Advanced in Control Engineering and Information Science
Avalable onlne at wwwscencedrectcom Avalable onlne at wwwscencedrectcom Proceda Proceda Engneerng Engneerng 00 (2011) 15000 000 (2011) 1642 1646 Proceda Engneerng wwwelsevercom/locate/proceda Advanced
More informationDynamic wetting property investigation of AFM tips in micro/nanoscale
Dynamc wettng property nvestgaton of AFM tps n mcro/nanoscale The wettng propertes of AFM probe tps are of concern n AFM tp related force measurement, fabrcaton, and manpulaton technques, such as dp-pen
More informationShape-adaptive DCT and Its Application in Region-based Image Coding
Internatonal Journal of Sgnal Processng, Image Processng and Pattern Recognton, pp.99-108 http://dx.do.org/10.14257/sp.2014.7.1.10 Shape-adaptve DCT and Its Applcaton n Regon-based Image Codng Yamn Zheng,
More informationX- 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 informationBackpropagation: 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 informationOptimal Workload-based Weighted Wavelet Synopses
Optmal Workload-based Weghted Wavelet Synopses Yoss Matas School of Computer Scence Tel Avv Unversty Tel Avv 69978, Israel matas@tau.ac.l Danel Urel School of Computer Scence Tel Avv Unversty Tel Avv 69978,
More informationModeling Inter-cluster and Intra-cluster Discrimination Among Triphones
Modelng Inter-cluster and Intra-cluster Dscrmnaton Among Trphones Tom Ko, Bran Mak and Dongpeng Chen Department of Computer Scence and Engneerng The Hong Kong Unversty of Scence and Technology Clear Water
More informationThree 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 informationOptimal Design of Nonlinear Fuzzy Model by Means of Independent Fuzzy Scatter Partition
Optmal Desgn of onlnear Fuzzy Model by Means of Independent Fuzzy Scatter Partton Keon-Jun Park, Hyung-Kl Kang and Yong-Kab Km *, Department of Informaton and Communcaton Engneerng, Wonkwang Unversty,
More informationMachine 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 informationSoftware Reliability Assessment Using High-Order Markov Chains
Internatonal Journal of Engneerng Scence Inventon ISSN (Onlne): 2319 6734, ISSN (Prnt): 2319 6726 www.jes.org Volume 3 Issue 7ǁ July 2014 ǁ PP.01-06 Software Relablty Assessment Usng Hgh-Order Markov Chans
More informationA New Feature of Uniformity of Image Texture Directions Coinciding with the Human Eyes Perception 1
A New Feature of Unformty of Image Texture Drectons Concdng wth the Human Eyes Percepton Xng-Jan He, De-Shuang Huang, Yue Zhang, Tat-Mng Lo 2, and Mchael R. Lyu 3 Intellgent Computng Lab, Insttute of Intellgent
More informationImplementation Naïve Bayes Algorithm for Student Classification Based on Graduation Status
Internatonal Journal of Appled Busness and Informaton Systems ISSN: 2597-8993 Vol 1, No 2, September 2017, pp. 6-12 6 Implementaton Naïve Bayes Algorthm for Student Classfcaton Based on Graduaton Status
More informationScheduling Remote Access to Scientific Instruments in Cyberinfrastructure for Education and Research
Schedulng Remote Access to Scentfc Instruments n Cybernfrastructure for Educaton and Research Je Yn 1, Junwe Cao 2,3,*, Yuexuan Wang 4, Lanchen Lu 1,3 and Cheng Wu 1,3 1 Natonal CIMS Engneerng and Research
More informationMachine 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 informationA New Token Allocation Algorithm for TCP Traffic in Diffserv Network
A New Token Allocaton Algorthm for TCP Traffc n Dffserv Network A New Token Allocaton Algorthm for TCP Traffc n Dffserv Network S. Sudha and N. Ammasagounden Natonal Insttute of Technology, Truchrappall,
More informationTHE FUZZY GROUP METHOD OF DATA HANDLING WITH FUZZY INPUTS. Yuriy Zaychenko
206 5 Knowledge Dalogue - Soluton THE FUZZY GROUP ETHOD OF DATA HANDLING WITH FUZZY INPUTS Yury Zaycheno Abstract: The problem of forecastng models constructng usng expermental data n terms of fuzzness,
More informationData Representation in Digital Design, a Single Conversion Equation and a Formal Languages Approach
Data Representaton n Dgtal Desgn, a Sngle Converson Equaton and a Formal Languages Approach Hassan Farhat Unversty of Nebraska at Omaha Abstract- In the study of data representaton n dgtal desgn and computer
More informationFuzzy Filtering Algorithms for Image Processing: Performance Evaluation of Various Approaches
Proceedngs of the Internatonal Conference on Cognton and Recognton Fuzzy Flterng Algorthms for Image Processng: Performance Evaluaton of Varous Approaches Rajoo Pandey and Umesh Ghanekar Department of
More informationRandom Variables and Probability Distributions
Random Varables and Probablty Dstrbutons Some Prelmnary Informaton Scales on Measurement IE231 - Lecture Notes 5 Mar 14, 2017 Nomnal scale: These are categorcal values that has no relatonshp of order or
More informationAn Ensemble Learning algorithm for Blind Signal Separation Problem
An Ensemble Learnng algorthm for Blnd Sgnal Separaton Problem Yan L 1 and Peng Wen 1 Department of Mathematcs and Computng, Faculty of Engneerng and Surveyng The Unversty of Southern Queensland, Queensland,
More informationMixed Linear System Estimation and Identification
48th IEEE Conference on Decson and Control, Shangha, Chna, December 2009 Mxed Lnear System Estmaton and Identfcaton A. Zymns S. Boyd D. Gornevsky Abstract We consder a mxed lnear system model, wth both
More informationGenerating Fuzzy Term Sets for Software Project Attributes using and Real Coded Genetic Algorithms
Generatng Fuzzy Ter Sets for Software Proect Attrbutes usng Fuzzy C-Means C and Real Coded Genetc Algorths Al Idr, Ph.D., ENSIAS, Rabat Alan Abran, Ph.D., ETS, Montreal Azeddne Zah, FST, Fes Internatonal
More informationOn Some Entertaining Applications of the Concept of Set in Computer Science Course
On Some Entertanng Applcatons of the Concept of Set n Computer Scence Course Krasmr Yordzhev *, Hrstna Kostadnova ** * Assocate Professor Krasmr Yordzhev, Ph.D., Faculty of Mathematcs and Natural Scences,
More information