Document Clustering Analysis Based on Hybrid PSO+K-means Algorithm

Size: px
Start display at page:

Download "Document Clustering Analysis Based on Hybrid PSO+K-means Algorithm"

Transcription

1 Document Clusterng Analyss Based on Hybrd PSO+K-means Algorthm Xaohu Cu, Thomas E. Potok Appled Software Engneerng Research Group, Computatonal Scences and Engneerng Dvson, Oak Rdge Natonal Laboratory, Oak Rdge, TN , USA cux, Abstract: There s a tremendous prolferaton n the amount of nformaton avalable on the largest shared nformaton source, the World Wde Web. Fast and hgh-qualty document clusterng algorthms play an mportant role n helpng users to effectvely navgate, summarze, and organze the nformaton. Recent studes have shown that parttonal clusterng algorthms are more sutable for clusterng large datasets. The K-means algorthm s the most commonly used parttonal clusterng algorthm because t can be easly mplemented and s the most effcent one n terms of the executon tme. The maor problem wth ths algorthm s that t s senstve to the selecton of the ntal partton and may converge to a local optma. In ths paper, we present a hybrd Partcle Swarm Optmzaton (PSO)+K-means document clusterng algorthm that performs fast document clusterng and can avod beng trapped n a local optmal soluton as well. For comparson purpose, we appled the PSO+K-means, PSO, K-means, and other two hybrd clusterng algorthms on four dfferent text document datasets. The number of documents n the datasets range from 204 to over 800, and the number of terms range from over 5000 to over The results llustrate that the PSO+K-means algorthm can generate the most compact clusterng results than other four algorthms. Keywords: Partcle Swarm Optmzaton, Text Dataset, Cluster Centrod, Vector Space Model INTRODUCTION Document clusterng s a fundamental operaton used n unsupervsed document organzaton, automatc topc extracton, and nformaton retreval. Clusterng nvolves dvdng a set of obects nto a specfed number of clusters [21]. The motvaton behnd clusterng a set of data s to fnd nherent structure n the data and expose ths structure as a set of groups. The data obects wthn each group should exhbt a large degree of smlarty whle the smlarty among dfferent clusters should be mnmzed [3, 9, 18]. There are two maor clusterng technques: Parttonng and Herarchcal [9]. Most document clusterng algorthms can be classfed nto these two groups. Herarchcal technques produce a nested sequence of partton, wth a sngle, all-nclusve cluster at the top and sngle clusters of ndvdual ponts at the bottom. The parttonng clusterng method seeks to partton a collecton of documents nto a set of non-overlappng groups, so as to maxmze the evaluaton value of clusterng. Although the herarchcal clusterng technque s often portrayed as a better qualty clusterng approach, ths technque does not contan any provson for the reallocaton of enttes, whch may have been poorly classfed n the early stages of the text analyss [9]. Moreover, the tme complexty of ths approach s quadratc [18]. In recent years, t has been recognzed that the parttonal clusterng technque s well suted for clusterng a large document dataset due to ther relatvely low computatonal requrements [18]. The tme complexty of the parttonng technque s almost lnear, whch makes t wdely used. The best-known parttonng clusterng algorthm s the K-means algorthm and ts varants [10]. Ths algorthm s smple, straghtforward and s based on the frm foundaton of analyss of varances. The K-means algorthm clusters a group of data vectors nto a predefned number of clusters. It starts wth a random ntal cluster center and keeps reassgnng the data obects n the dataset to cluster centers based on the smlarty between the data obect and the cluster center. The reassgnment procedure wll not stop untl

2 a convergence crteron s met (e.g., the fxed teraton number or the cluster result does not change after a certan number of teratons). The man drawback of the K-means algorthm s that the cluster result s senstve to the selecton of the ntal cluster centrods and may converge to the local optma [16]. Therefore, the ntal selecton of the cluster centrods decdes the man processng of K- means and the partton result of the dataset as well. The man processng of K-means s to search the local optmal soluton n the vcnty of the ntal soluton and to refne the partton result. The same ntal cluster centrods n a dataset wll always generate the same cluster results. However, f good ntal clusterng centrods can be obtaned usng any of the other technques, the K-means would work well n refnng the clusterng centrods to fnd the optmal clusterng centers [2]. It s necessary to employee some other global optmal searchng algorthm for generatng ths ntal cluster centrods. The Partcle Swarm Optmzaton (PSO) algorthm s a populaton based stochastc optmzaton technque that can be used to fnd an optmal, or near optmal, soluton to a numercal and qualtatve problem [4, 11, 17]. The PSO algorthm can be used to generate good ntal cluster centrods for the K-means. In ths paper, we present a hybrd PSO+K-means document clusterng algorthm that performs fast document clusterng and can avod beng trapped n a local optmal soluton. The results from our experments ndcate that the PSO+K-means algorthm can generate the best results n ust 50 teratons n comparson wth the K-means algorthm and the PSO algorthm. The remander of ths paper s organzed as follows: Secton 2 provdes the methods of representng documents n clusterng algorthms and of computng the smlarty between documents. Secton 3 provdes a general overvew of the PSO algorthm. The hybrd PSO+K-means clusterng algorthms are descrbed n Secton 4. Secton 5 provdes the detaled expermental setup and results for comparng the performance of the PSO+K-means algorthm wth the K-means, PSO, and other hybrd approaches. The dscusson of the experment s results s also presented. The concluson s n Secton 6. PRELIMINARIES Document representaton: In most clusterng algorthms, the dataset to be clustered s represented as a set of vectors X={x 1, x 2,., x n }, where the vector x corresponds to a sngle obect and s called the feature vector. The feature vector should nclude proper features to represent the obect. The text document obects can be represented usng the Vector Space Model (VSM) [8]. In ths model, the content of a document s formalzed as a dot n the multdmensonal space and represented by a vector d, such as d= { w 1, w2,... wn}, where w ( = 1,2,,n) s the term weght of the term t n one document. The term weght value represents the sgnfcance of ths term n a document. To calculate the term weght, the occurrence frequency of the term wthn a document and n the entre set of documents must be consdered. The most wdely used weghtng scheme combnes the Term Frequency wth Inverse Document Frequency (TF-IDF) [8]. The weght of term n document s gven n equaton 1: w = tf df = tf *log ( n / df ) (1) * 2 where tf s the number of occurrences of term n the document ; df ndcates the term frequency n the collectons of documents; and n s the total number of documents n the collecton. Ths weghtng scheme dscounts the frequent words wth lttle dscrmnatng power. The smlarty metrc: The smlarty between two documents needs to be measured n a clusterng analyss. Over the years, two promnent ways have been proposed to compute the smlarty between documents m p and m. The frst method s based on Mnkowsk dstances [5], gven by: D ( m, m ) n p d m = = 1 m n, p m, 1/ n (2) For n =2, we obtan the Eucldean dstance. In order to manpulate equvalent threshold dstances, consderng that the dstance ranges wll vary accordng to the dmenson number, ths algorthm uses the normalzed Eucldean dstance as the smlarty metrc of two documents, m p and m, n the vector space. Equaton 3 represents the dstance measurement formula: d m d( m, m ) = ( m m ) 2 / d (3) p k = 1 where m p and m are two document vectors; d m denotes the dmenson number of the vector space; m pk and m k stand for the documents m p and m s weght values n dmenson k. pk k m

3 The other commonly used smlarty measure n document clusterng s the cosne correlaton measure [15], gven by: t mpm cos( m p, m ) = (4) m m p where m t m p denotes the dot-product of the two document vectors;. ndcates the length of the vector. Both smlarty metrcs are wdely used n the text document clusterng lteratures. BACKGROUND OF THE PSO ALGORITHM PSO was orgnally developed by Eberhart and Kennedy n 1995 [11], and was nspred by the socal behavor of a brd flock. In the PSO algorthm, the brds n a flock are symbolcally represented as partcles. These partcles can be consdered as smple agents flyng through a problem space. A partcle s locaton n the mult-dmensonal problem space represents one soluton for the problem. When a partcle moves to a new locaton, a dfferent problem soluton s generated. Ths soluton s evaluated by a ftness functon that provdes a quanttatve value of the soluton s utlty. The velocty and drecton of each partcle movng along each dmenson of the problem space wll be altered wth each generaton of movement. In combnaton, the partcle s personal experence, P d and ts neghbors experence, P gd nfluence the movement of each partcle through a problem space. The random values, rand 1 and rand 2, are used for the sake of completeness, that s, to make sure that partcles explore wde search space before convergng around the optmal soluton. The values of c 1 and c 2 control the weght balance of P d and P gd n decdng the partcle s next movement velocty. For every generaton, the partcle s new locaton s computed by addng the partcle s current velocty, V-vector, to ts locaton, X-vector. Mathematcally, gven a multdmensonal problem space, the th partcle changes ts velocty and locaton accordng to the followng equatons [11]: vd = w* vd + c * rand1 *( pd xd ) + c2 * rand2 *( p x = x + v d d x 1 gd d d ) (5a) (5b) where w denotes the nerta weght factor; p d s the locaton of the partcle that experences the best ftness value; p gd s the locaton of the partcles that experence a global best ftness value; c 1 and c 2 are constants and are known as acceleraton coeffcents; d denotes the dmenson of the problem space; rand 1, rand 2 are random values n the range of (0, 1). The nerta weght factor w provdes the necessary dversty to the swarm by changng the momentum of partcles to avod the stagnaton of partcles at the local optma. The emprcal research conducted by Eberhart and Sh [7] shows mprovement of search effcency through gradually decreasng the value of nerta weght factor from a hgh value durng the search. Equaton 5a requres each partcle to record ts current coordnate X d, ts velocty V d that ndcates the speed of ts movement along the dmensons n a problem space, and the coordnates P d and P gd where the best ftness values were computed. The best ftness values are updated at each generaton, based on equaton 6, P ( t) P ( t + 1) = X ( t + 1) f ( X ( t + 1)) f ( X ( t)) (6) f ( X ( t + 1)) > f ( X ( t)) where the symbol f denotes the ftness functon; P (t) stands for the best ftness values and the coordnaton where the value was calculated; and t denotes the generaton step. It s possble to vew the clusterng problem as an optmzaton problem that locates the optmal centrods of the clusters rather than fndng an optmal partton. Ths vew offers us a chance to apply PSO optmal algorthm on the clusterng soluton. In [6], we proposed a PSO document clusterng algorthm. Contrary to the localzed searchng n the K-means algorthm, the PSO clusterng algorthm performs a globalzed search n the entre soluton space [4, 17]. Utlzng the PSO algorthm s optmal ablty, f gven enough tme, the PSO clusterng algorthm we proposed could generate more compact clusterng results from the document datasets than the tradtonal K-means clusterng algorthm. However, n order to cluster the large document datasets, PSO requres much more teraton (generally more than 500 teratons) to converge to the optma than the K-mean algorthm does. Although the PSO algorthm s nherently parallel and can be mplemented usng parallel hardware, such as a computer cluster, the computaton requrement for clusterng extremely huge document datasets s stll hgh. In terms of executon tme, the K-means algorthm s the most effcent for the large dataset [1]. The K-means algorthm tends to converge faster than the PSO, but t usually only fnds the local maxmum. Therefore, we face a dlemma regardng choosng the algorthm for clusterng a large

4 document dataset. Base on ths reason, we proposed a hybrd PSO+K-means document clusterng algorthm. HYBRID PSO+K-MEANS ALGORITHM In the hybrd PSO+K-means algorthm, the multdmensonal document vector space s modeled as a problem space. Each term n the document dataset represents one dmenson of the problem space. Each document vector can be represented as a dot n the problem space. The whole document dataset can be represented as a multple dmenson space wth a large number of dots n the space. The hybrd PSO+Kmeans algorthm ncludes two modules, the PSO module and K-means module. At the ntal stage, the PSO module s executed for a short perod to search for the clusters centrod locatons. The locatons are transferred to the K-means module for refnng and generatng the fnal optmal clusterng soluton. The PSO module: A sngle partcle n the swarm represents one possble soluton for clusterng the document collecton. Therefore, a swarm represents a number of canddate clusterng solutons for the document collecton. Each partcle mantans a matrx X = (C 1, C 2,, C,.., C k ), where C represents the th cluster centrod vector and k s the cluster number. At each teraton, the partcle adusts the centrod vector poston n the vector space accordng to ts own experence and those of ts neghbors. The average dstance between a cluster centrod and a document s used as the ftness value to evaluate the soluton represented by each partcle. The ftness value s measured by the equaton below: f p Nc = 1 { = = 1 d( o, m c p N ) } (7) where m denotes the th document vector, whch belongs to cluster ; O s the centrod vector of th cluster; d(o, m ) s the dstance between document m and the cluster centrod O. ; P stands for the document number, whch belongs to cluster C ; N c stands for the cluster number. The PSO module can be summarzed as: (1) At the ntal stage, each partcle randomly chooses k numbers of document vectors from the document collecton as the cluster centrod vectors. (2) For each partcle: (a) Assgnng each document vector n the document set to the closest centrod vector. (b) Calculatng the ftness value based on equaton 7. (c) Usng the velocty and partcle poston to update equatons 5a and 5b and to generate the next solutons. (3) Repeatng step (2) untl one of followng termnaton condtons s satsfed. (a) The maxmum number of teratons s exceeded or (b) The average change n centrod vectors s less than a predefned value. The K-means module: The K-means module wll nhert the PSO module s result as the ntal clusterng centrods and wll contnue processng the optmal centrods to generate the fnal result. The K-means module can be summarzed as: (1) Inhertng cluster centrod vectors from the PSO module. (2) Assgnng each document vector to the closest cluster centrods. (3) Recalculatng the cluster centrod vector c usng equaton 8. c 1 = n d d S (8) where d denotes the document vectors that belong to cluster S ; c stands for the centrod vector; n s the number of document vectors belong to cluster S. (4) Repeatng step 2 and 3 untl the convergence s acheved. In the PSO+K-means algorthm, the ablty of globalzed searchng of the PSO algorthm and the fast convergence of the K-means algorthm are combned. The PSO algorthm s used at the ntal stage to help dscoverng the vcnty of the optmal soluton by a global search. The result from PSO s used as the ntal seed of the K-means algorthm, whch s appled for refnng and generatng the fnal result. EXPERIMENTS AND RESULTS Datasets: We used four dfferent document collectons to compare the performance of the K-

5 means, PSO and hybrd PSO+K-means algorthms wth dfferent combnaton models. These document datasets are derved from the TREC-5, TREC-6, and TREC-7 collectons [19]. A descrpton of the test datasets s gven n Table 1. In those document datasets, the very common words (e.g. functon words: a, the, n, to ; pronouns: I, he, she, t ) are strpped out completely and dfferent forms of a word are reduced to one canoncal form by usng Porter s algorthm [13]. In order to reduce the mpact of the length varatons of dfferent documents, each document vector s normalzed so that t s of unt length. The document number n each dataset ranges from 204 to 878. The term numbers of each dataset are all over Table 1: Summary of text document datasets Data Number of Number of Number of documents terms classes Dataset Dataset Dataset Dataset Expermental setup: The K-means, PSO and PSO+K-means clusterng approaches are appled on the four datasets, respectvely. The Eucldan dstance measure and cosne correlaton measure are used as the smlarty metrcs n each algorthm, respectvely. Some researchers [12, 20] proposed usng the K- means+pso hybrd algorthm for clusterng low dmenson datasets. They argue that the K-means algorthm tends to converge faster than other clusterng algorthms, but usually wth a less accurate clusterng. The performance of the clusterng algorthm can be mproved by seedng the ntal swarm wth the result of the K-means algorthm. To compare the performance of dfferent knds of hybrd algorthms, we appled K-means+PSO and K-means+PSO+Kmeans algorthms on the four datasets, respectvely. We have notced that K-means clusterng algorthms can converge to a stable soluton wthn 20 teratons when appled to most document datasets. The PSO usually needs to repeat for more than 100 teratons to generate a stable soluton. For an easy comparson, the K-means and PSO approaches run 50 teratons. In the K-means+PSO approach, the K-means algorthm s frst executed for 25 teratons. The result of the K- means algorthm s then used as the ntal cluster centrod n the PSO algorthm, and the PSO algorthm executes for another 25 teratons to generate the fnal result. The PSO+k-means approach has the same executon procedure, except that t frst executes the PSO algorthm for 25 teratons and uses the PSO result as the ntal seed for the K-means algorthm. In the K-means+PSO+K-means approach, the K-means algorthm s frst executed for 25 teratons. The result of the K-means algorthm s then used as the ntal cluster centrod n the PSO algorthm and the PSO algorthm executes for 25 teratons, the global best soluton result from the PSO algorthm s used as the ntal cluster centrod of the K-means algorthm, and the K-means algorthm executes for another 25 teratons to generate the fnal result. In these fve dfferent algorthms, the total executng teraton number for K-means, PSO, K-means+PSO, and PSO+K-means s 50. The total executng number for the K-means+PSO+K-means teraton approach s 75. No parameter needs to be set up for the K-means algorthm. In the PSO clusterng algorthm, because of the extremely hgh dmensonal soluton space of the text document datasets, we choose 50 partcles for all the PSO algorthms nstead of choosng 20 to 30 partcles recommended n [4, 17]. In the PSO algorthm, the nerta weght w s ntally set as 0.72 and the acceleraton coeffcent constants c1 and c2 are set as These values are chosen based on the results of [17]. In the PSO algorthm, the nerta weght wll reduce 1% n value at each teraton to ensure good convergence. However, the nerta weght n all hybrd algorthms s kept constant to ensure a globalzed search. Results: The ftness equaton 7 s used not only n the PSO algorthm for the ftness value calculaton, but also n the evaluaton of the cluster qualty. It ndcates the value of the average dstance between documents and the cluster centrod to whch they belong (ADVDC). The smaller the ADVDC value, the more compact the clusterng soluton s. Table 2 demonstrates the expermental results by usng the K- means, PSO, K-means+PSO, PSO+K-means and K- means+pso+k-means respectvely. Ten smulatons are performed separately. The average ADVDC values and standard dvson are recorded n Table 2. To llustrate the convergence behavor of dfferent clusterng algorthms, the clusterng ADVDC values at each teraton are recorded when these fve algorthms are appled on datasets separately. As shown n Table 2, the PSO+K-means hybrd clusterng approach generates the clusterng result that has the lowest ADVDC value for all four datasets usng the Eucldan smlarty metrc and the Cosne

6 Table 2: Performance comparson of K-means, PSO, K-means+PSO, PSO+K-means and K-means+PSO+K-means Dataset 1 Dataset 2 Dataset 3 Dataset 4 ADVDC value K-means PSO PSO+K-means K-means+PSO K-means+PSO +K-means Eucldan 8.238± ± ± ± ±0.138 Cosne 8.999± ± ± ± ±0.231 Eucldan 7.245± ± ± ± ±0.086 Cosne 8.074± ± ± ± ±0.152 Eucldan 4.788± ± ± ± ±0.247 Cosne 5.093± ± ± ± ±0.267 Eucldan 9.09± ± ± ± ±0.363 Cosne 10.22± ± ± ± ±0.343 correlaton smlarty metrc. The results from the PSO approach have mprovements compared to the results of the K-means approach when usng the Eucldan smlarty metrc. However, when the smlarty metrc s changed wth the cosne correlaton metrc, the K- means algorthm has a better performance than the PSO algorthm. The K-means+PSO and K- means+pso+k-means approaches do not have sgnfcant mprovements compared to the result of the K-means approach. Fgure 1 llustrates the convergence behavors of these algorthms on the document dataset 1 usng the Eucldan dstance as a smlarty metrc. In Fgure 1, the K-means algorthm converges quckly but prematurely wth hgh quantzaton error. As shown n fgure 1, the ADVDC value of the K-means algorthm s sharply reduced from 11.3 to 8.2 wthn 10 teratons and fxed at 8.2. In Fgure 1, t s hard to separate the curve lnes that represent the K-means+PSO and K- means+pso+k-means approaches from the K-means approach. The three lnes nearly overlap each other, whch ndcates these three algorthms have nearly the same convergence behavor. The PSO approach s ADVDC value s quckly converged from 11.3 to 6.7 wthn 30 teratons. The reducton of the ADVDC value n PSO s not as sharp as n K-means and becomes smoothly after 30 teratons. The curvy lne s tendency ndcates that f more teratons are executed, the dstance average value may reduce further although the reducton speed wll be very slow. The PSO+K-means approach s performance sgnfcantly mproves. In the frst 25 teratons, the PSO+K-means algorthm has smlar convergence behavor because wthn 1 to 25 teratons, the PSO and the PSO+K-means algorthms execute the same PSO optmal code. After 25 teratons, the ADVDC value has a sharp reducton wth the value reduced from 6.7 to 4.7 and mantans a stable value wthn 10 teratons. Dscusson: Usng hybrd algorthms for boostng the clusterng performance s not a novel dea. However, most of hybrd algorthms use K-means algorthm for generatng the ntal clusterng seeds for other optmal algorthms. To the best of the author s knowledge, there no hybrd algorthm that uses PSO optmal algorthm generatng ntal seed for K-means clusterng. In [20], Merwe and Engelbrecht argued that the performance of the PSO clusterng algorthm could be mproved by seedng the ntal swarm wth the result of the K-means algorthm. They conducted smulatons on some low dmenson datasets wth 10 partcles and 1000 teratons. However, from the expermental results n Table 2, we notced the K- means +PSO algorthm does not show any mprovement n the large document datasets. The dfferences between the experments n the present research and the experment n [20] are (a) the datasets used here are document data wth more than 5000 terms, whch s also the dmensonal number of the PSO searchng space; (b) The teratons runnng n each smulaton are no more than 75. In the K- means+pso approach, the K-means algorthm generates a local optma result, whch s used as the ntal status of one partcle n the PSO algorthm. The hgh ftness value of ths ntal partcle s value wll drectly force the partcle to start the optmal refnng stage. Although other partcles are randomly deployed n the searchng space, the hgh average dstance value

7 ADVDC K-means PSO PSO+K K+PSO K+PSO+K Iteraton Fgure 1: The convergence behavors of dfferent clusterng algorthm (PSO+K: the PSO+K-means Algorthm, K+PSO: the K-means + PSO algorthm, K+PSO+K: the K-means +PSO+K-means algorthm) of the ntal soluton that the PSO algorthm nherted from the K-means algorthm wll attract other partcles to converge quckly nto the vcnty of the local optma. The performance results n Table 2 and the convergence behavors of the K-means+PSO approach n Fgure 1 llustrate ths. The PSO+K-means algorthm generates the hghest clusterng compact result n the experments. The average dstance value s the lowest. In the PSO+Kmeans algorthm clusterng experment, although 25 teratons s not enough for the PSO to dscover the optmal soluton, t has a hgh possblty that one partcle s soluton s located n the vcnty of the global soluton. The result of the PSO s used as the ntal seed of the K-means algorthm and the K-means algorthm can quckly locate the optma wth a low dstance average value. Comparson of the performance of these fve approaches n our experments llustrates that the sequence of hybrd K- means algorthms s very mportant. CONCLUSION In ths study, we presented a document clusterng algorthm, the PSO+K-means algorthm, whch can be regarded as a hybrd of the PSO and K-means algorthms. In the general PSO algorthm, PSO can conduct a globalzed searchng for the optmal clusterng, but requres more teraton numbers and computaton than the K-means algorthm does. The K- means algorthm tends to converge faster than the PSO algorthm, but usually can be trapped n a local optmal area. The PSO+K-means algorthm combnes the ablty of the globalzed searchng of the PSO algorthm and the fast convergence of the K-means algorthm and can avod the drawback of both algorthms. The algorthm ncludes two modules, the PSO module and the K-means module. The PSO module s executed for a short perod at the ntal stage to dscover the vcnty of the optmal soluton by a global search and at the same tme to avod consumng hgh computaton. The result from the PSO module s used as the ntal seed of the K-means module. The K-means algorthm wll be appled for refnng and generatng the fnal result. Our expermental results llustrate that usng ths hybrd PSO+K-means algorthm can generate hgher compact clusterng than usng ether PSO or K-means alone. The results from the three dfferent hybrd K-means algorthms llustrates that performng the K-means n advance of the PSO module n the hybrd algorthm wll reduce the globalzed searchng ablty of the PSO algorthm and lower the whole algorthm s performance.

8 ACKNOWLEDGMENT Oak Rdge Natonal Laboratory s managed by UT- Battelle LLC for the U.S. Department of Energy under contract number DE-AC05_00OR REFERENCES 1. Al-Sultan, K. S. and Khan, M. M Computatonal experence on four algorthms for the hard clusterng problem. Pattern Recogn. Lett. 17, 3, Anderberg, M. R., Cluster Analyss for Applcatons. Academc Press, Inc., New York, NY. 3. Berkhn, P., Survey of clusterng data mnng technques. Accrue Software Research Paper. 4. Carlsle, A. and Dozer, G., An Off-The- Shelf PSO, Proceedngs of the 2001 Workshop on Partcle Swarm Optmzaton, pp. 1-6, Indanapols, IN 5. Cos K., Pedrycs W., Swnarsk R., Data Mnng Methods for Knowledge Dscovery, Kluwer Academc Publshers. 6. Cu X., Potok T. E., Document Clusterng usng Partcle Swarm Optmzaton, IEEE Swarm Intellgence Symposum 2005, Pasadena, Calforna. 7. Eberhart, R.C., and Sh, Y., Comparng Inerta Weghts and Constrcton Factors n Partcle Swarm Optmzaton, 2000 Congress on Evolutonary Computng, vol. 1, pp Evertt, B., Cluster Analyss. 2 nd Edton. Halsted Press, New York. 9. Jan A. K., Murty M. N., and Flynn P. J., Data Clusterng: A Revew, ACM Computng Survey, Vol. 31, No. 3, pp Hartgan, J. A Clusterng Algorthms. John Wley and Sons, Inc., New York, NY. 11. Kennedy J., Eberhart R. C. and Sh Y., Swarm Intellgence, Morgan Kaufmann, New York. 12. Omran, M., Salman, A. and Engelbrecht, A. P., Image classfcaton usng partcle swarm optmzaton. Proceedngs of the 4th Asa-Pacfc Conference on Smulated Evoluton and Learnng 2002 (SEAL 2002), Sngapore. pp Porter, M.F., An Algorthm for Suffx Strppng. Program, 14 no. 3, pp Salton G., Automatc Text Processng. Addson-Wesley. 15. Salton G. and Buckley C., Term-weghtng approaches n automatc text retreval. Informaton Processng and Management, 24 (5): pp Selm, S. Z. And Ismal, M. A K-means type algorthms: A generalzed convergence theorem and characterzaton of local optmalty. IEEE Trans. Pattern Anal. Mach. Intell. 6, Sh, Y. H., Eberhart, R. C., Parameter Selecton n Partcle Swarm Optmzaton, The 7th Annual Conference on Evolutonary Programmng, San Dego, CA. 18. Stenbach M., Karyps G., Kumar V., A Comparson of Document Clusterng Technques. TextMnng Workshop, KDD. 19. TREC Text Retreval Conference Van D. M., Engelbrecht, A. P., Data clusterng usng partcle swarm optmzaton. Proceedngs of IEEE Congress on Evolutonary Computaton 2003 (CEC 2003), Canbella, Australa. pp Zhao Y. and Karyps G., Emprcal and Theoretcal Comparsons of Selected Crteron Functons for Document Clusterng, Machne Learnng, 55 (3): pp

The Development of Web Log Mining Based on Improve-K-Means Clustering Analysis

The Development of Web Log Mining Based on Improve-K-Means Clustering Analysis The Development of Web Log Mnng Based on Improve-K-Means Clusterng Analyss TngZhong Wang * College of Informaton Technology, Luoyang Normal Unversty, Luoyang, 471022, Chna wangtngzhong2@sna.cn Abstract.

More information

Descriptive Models. Cluster Analysis. Example. General Applications of Clustering. Examples of Clustering Applications

Descriptive Models. Cluster Analysis. Example. General Applications of Clustering. Examples of Clustering Applications CMSC828G Prncples of Data Mnng Lecture #9 Today s Readng: HMS, chapter 9 Today s Lecture: Descrptve Modelng Clusterng Algorthms Descrptve Models model presents the man features of the data, a global summary

More information

A DATA MINING APPLICATION IN A STUDENT DATABASE

A DATA MINING APPLICATION IN A STUDENT DATABASE JOURNAL OF AERONAUTICS AND SPACE TECHNOLOGIES JULY 005 VOLUME NUMBER (53-57) A DATA MINING APPLICATION IN A STUDENT DATABASE Şenol Zafer ERDOĞAN Maltepe Ünversty Faculty of Engneerng Büyükbakkalköy-Istanbul

More information

An Interest-Oriented Network Evolution Mechanism for Online Communities

An Interest-Oriented Network Evolution Mechanism for Online Communities An Interest-Orented Network Evoluton Mechansm for Onlne Communtes Cahong Sun and Xaopng Yang School of Informaton, Renmn Unversty of Chna, Bejng 100872, P.R. Chna {chsun,yang}@ruc.edu.cn Abstract. Onlne

More information

A hybrid global optimization algorithm based on parallel chaos optimization and outlook algorithm

A hybrid global optimization algorithm based on parallel chaos optimization and outlook algorithm Avalable onlne www.ocpr.com Journal of Chemcal and Pharmaceutcal Research, 2014, 6(7):1884-1889 Research Artcle ISSN : 0975-7384 CODEN(USA) : JCPRC5 A hybrd global optmzaton algorthm based on parallel

More information

320 The Internatonal Arab Journal of Informaton Technology, Vol. 5, No. 3, July 2008 Comparsons Between Data Clusterng Algorthms Osama Abu Abbas Computer Scence Department, Yarmouk Unversty, Jordan Abstract:

More information

Module 2 LOSSLESS IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur

Module 2 LOSSLESS IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur Module LOSSLESS IMAGE COMPRESSION SYSTEMS Lesson 3 Lossless Compresson: Huffman Codng Instructonal Objectves At the end of ths lesson, the students should be able to:. Defne and measure source entropy..

More information

DEFINING %COMPLETE IN MICROSOFT PROJECT

DEFINING %COMPLETE IN MICROSOFT PROJECT CelersSystems DEFINING %COMPLETE IN MICROSOFT PROJECT PREPARED BY James E Aksel, PMP, PMI-SP, MVP For Addtonal Informaton about Earned Value Management Systems and reportng, please contact: CelersSystems,

More information

An Alternative Way to Measure Private Equity Performance

An Alternative Way to Measure Private Equity Performance An Alternatve Way to Measure Prvate Equty Performance Peter Todd Parlux Investment Technology LLC Summary Internal Rate of Return (IRR) s probably the most common way to measure the performance of prvate

More information

Forecasting the Direction and Strength of Stock Market Movement

Forecasting the Direction and Strength of Stock Market Movement Forecastng the Drecton and Strength of Stock Market Movement Jngwe Chen Mng Chen Nan Ye cjngwe@stanford.edu mchen5@stanford.edu nanye@stanford.edu Abstract - Stock market s one of the most complcated systems

More information

Face Verification Problem. Face Recognition Problem. Application: Access Control. Biometric Authentication. Face Verification (1:1 matching)

Face Verification Problem. Face Recognition Problem. Application: Access Control. Biometric Authentication. Face Verification (1:1 matching) Face Recognton Problem Face Verfcaton Problem Face Verfcaton (1:1 matchng) Querymage face query Face Recognton (1:N matchng) database Applcaton: Access Control www.vsage.com www.vsoncs.com Bometrc Authentcaton

More information

Survey on Virtual Machine Placement Techniques in Cloud Computing Environment

Survey on Virtual Machine Placement Techniques in Cloud Computing Environment Survey on Vrtual Machne Placement Technques n Cloud Computng Envronment Rajeev Kumar Gupta and R. K. Paterya Department of Computer Scence & Engneerng, MANIT, Bhopal, Inda ABSTRACT In tradtonal data center

More information

On the Optimal Control of a Cascade of Hydro-Electric Power Stations

On the Optimal Control of a Cascade of Hydro-Electric Power Stations On the Optmal Control of a Cascade of Hydro-Electrc Power Statons M.C.M. Guedes a, A.F. Rbero a, G.V. Smrnov b and S. Vlela c a Department of Mathematcs, School of Scences, Unversty of Porto, Portugal;

More information

Cluster Analysis of Data Points using Partitioning and Probabilistic Model-based Algorithms

Cluster Analysis of Data Points using Partitioning and Probabilistic Model-based Algorithms Internatonal Journal of Appled Informaton Systems (IJAIS) ISSN : 2249-0868 Foundaton of Computer Scence FCS, New York, USA Volume 7 No.7, August 2014 www.jas.org Cluster Analyss of Data Ponts usng Parttonng

More information

Feature selection for intrusion detection. Slobodan Petrović NISlab, Gjøvik University College

Feature selection for intrusion detection. Slobodan Petrović NISlab, Gjøvik University College Feature selecton for ntruson detecton Slobodan Petrovć NISlab, Gjøvk Unversty College Contents The feature selecton problem Intruson detecton Traffc features relevant for IDS The CFS measure The mrmr measure

More information

"Research Note" APPLICATION OF CHARGE SIMULATION METHOD TO ELECTRIC FIELD CALCULATION IN THE POWER CABLES *

Research Note APPLICATION OF CHARGE SIMULATION METHOD TO ELECTRIC FIELD CALCULATION IN THE POWER CABLES * Iranan Journal of Scence & Technology, Transacton B, Engneerng, ol. 30, No. B6, 789-794 rnted n The Islamc Republc of Iran, 006 Shraz Unversty "Research Note" ALICATION OF CHARGE SIMULATION METHOD TO ELECTRIC

More information

A COPMARISON OF PARTICLE SWARM OPTIMIZATION AND THE GENETIC ALGORITHM

A COPMARISON OF PARTICLE SWARM OPTIMIZATION AND THE GENETIC ALGORITHM A COPMARISON OF PARTICLE SWARM OPTIMIZATION AND THE GENETIC ALGORITHM Rana Hassan * Babak Cohanm Olver de Weck Massachusetts Insttute of Technology, Cambrdge, MA, 39 Gerhard Venter Vanderplaats Research

More information

benefit is 2, paid if the policyholder dies within the year, and probability of death within the year is ).

benefit is 2, paid if the policyholder dies within the year, and probability of death within the year is ). REVIEW OF RISK MANAGEMENT CONCEPTS LOSS DISTRIBUTIONS AND INSURANCE Loss and nsurance: When someone s subject to the rsk of ncurrng a fnancal loss, the loss s generally modeled usng a random varable or

More information

APPLICATION OF PROBE DATA COLLECTED VIA INFRARED BEACONS TO TRAFFIC MANEGEMENT

APPLICATION OF PROBE DATA COLLECTED VIA INFRARED BEACONS TO TRAFFIC MANEGEMENT APPLICATION OF PROBE DATA COLLECTED VIA INFRARED BEACONS TO TRAFFIC MANEGEMENT Toshhko Oda (1), Kochro Iwaoka (2) (1), (2) Infrastructure Systems Busness Unt, Panasonc System Networks Co., Ltd. Saedo-cho

More information

Product Quality and Safety Incident Information Tracking Based on Web

Product Quality and Safety Incident Information Tracking Based on Web Product Qualty and Safety Incdent Informaton Trackng Based on Web News 1 Yuexang Yang, 2 Correspondng Author Yyang Wang, 2 Shan Yu, 2 Jng Q, 1 Hual Ca 1 Chna Natonal Insttute of Standardzaton, Beng 100088,

More information

Damage detection in composite laminates using coin-tap method

Damage detection in composite laminates using coin-tap method Damage detecton n composte lamnates usng con-tap method S.J. Km Korea Aerospace Research Insttute, 45 Eoeun-Dong, Youseong-Gu, 35-333 Daejeon, Republc of Korea yaeln@kar.re.kr 45 The con-tap test has the

More information

A Simple Approach to Clustering in Excel

A Simple Approach to Clustering in Excel A Smple Approach to Clusterng n Excel Aravnd H Center for Computatonal Engneerng and Networng Amrta Vshwa Vdyapeetham, Combatore, Inda C Rajgopal Center for Computatonal Engneerng and Networng Amrta Vshwa

More information

Vision Mouse. Saurabh Sarkar a* University of Cincinnati, Cincinnati, USA ABSTRACT 1. INTRODUCTION

Vision Mouse. Saurabh Sarkar a* University of Cincinnati, Cincinnati, USA ABSTRACT 1. INTRODUCTION Vson Mouse Saurabh Sarkar a* a Unversty of Cncnnat, Cncnnat, USA ABSTRACT The report dscusses a vson based approach towards trackng of eyes and fngers. The report descrbes the process of locatng the possble

More information

Cluster Analysis. Cluster Analysis

Cluster Analysis. Cluster Analysis Cluster Analyss Cluster Analyss What s Cluster Analyss? Types of Data n Cluster Analyss A Categorzaton of Maor Clusterng Methos Parttonng Methos Herarchcal Methos Densty-Base Methos Gr-Base Methos Moel-Base

More information

L10: Linear discriminants analysis

L10: Linear discriminants analysis L0: Lnear dscrmnants analyss Lnear dscrmnant analyss, two classes Lnear dscrmnant analyss, C classes LDA vs. PCA Lmtatons of LDA Varants of LDA Other dmensonalty reducton methods CSCE 666 Pattern Analyss

More information

Ants Can Schedule Software Projects

Ants Can Schedule Software Projects Ants Can Schedule Software Proects Broderck Crawford 1,2, Rcardo Soto 1,3, Frankln Johnson 4, and Erc Monfroy 5 1 Pontfca Unversdad Católca de Valparaíso, Chle FrstName.Name@ucv.cl 2 Unversdad Fns Terrae,

More information

Conversion between the vector and raster data structures using Fuzzy Geographical Entities

Conversion between the vector and raster data structures using Fuzzy Geographical Entities Converson between the vector and raster data structures usng Fuzzy Geographcal Enttes Cdála Fonte Department of Mathematcs Faculty of Scences and Technology Unversty of Combra, Apartado 38, 3 454 Combra,

More information

Patterns Antennas Arrays Synthesis Based on Adaptive Particle Swarm Optimization and Genetic Algorithms

Patterns Antennas Arrays Synthesis Based on Adaptive Particle Swarm Optimization and Genetic Algorithms IJCSI Internatonal Journal of Computer Scence Issues, Vol. 1, Issue 1, No 2, January 213 ISSN (Prnt): 1694-784 ISSN (Onlne): 1694-814 www.ijcsi.org 21 Patterns Antennas Arrays Synthess Based on Adaptve

More information

A New Task Scheduling Algorithm Based on Improved Genetic Algorithm

A New Task Scheduling Algorithm Based on Improved Genetic Algorithm A New Task Schedulng Algorthm Based on Improved Genetc Algorthm n Cloud Computng Envronment Congcong Xong, Long Feng, Lxan Chen A New Task Schedulng Algorthm Based on Improved Genetc Algorthm n Cloud Computng

More information

Course outline. Financial Time Series Analysis. Overview. Data analysis. Predictive signal. Trading strategy

Course outline. Financial Time Series Analysis. Overview. Data analysis. Predictive signal. Trading strategy Fnancal Tme Seres Analyss Patrck McSharry patrck@mcsharry.net www.mcsharry.net Trnty Term 2014 Mathematcal Insttute Unversty of Oxford Course outlne 1. Data analyss, probablty, correlatons, vsualsaton

More information

A Secure Password-Authenticated Key Agreement Using Smart Cards

A Secure Password-Authenticated Key Agreement Using Smart Cards A Secure Password-Authentcated Key Agreement Usng Smart Cards Ka Chan 1, Wen-Chung Kuo 2 and Jn-Chou Cheng 3 1 Department of Computer and Informaton Scence, R.O.C. Mltary Academy, Kaohsung 83059, Tawan,

More information

A Comparative Study of Data Clustering Techniques

A Comparative Study of Data Clustering Techniques A COMPARATIVE STUDY OF DATA CLUSTERING TECHNIQUES A Comparatve Study of Data Clusterng Technques Khaled Hammouda Prof. Fakhreddne Karray Unversty of Waterloo, Ontaro, Canada Abstract Data clusterng s a

More information

An Enhanced Super-Resolution System with Improved Image Registration, Automatic Image Selection, and Image Enhancement

An Enhanced Super-Resolution System with Improved Image Registration, Automatic Image Selection, and Image Enhancement An Enhanced Super-Resoluton System wth Improved Image Regstraton, Automatc Image Selecton, and Image Enhancement Yu-Chuan Kuo ( ), Chen-Yu Chen ( ), and Chou-Shann Fuh ( ) Department of Computer Scence

More information

Gender Classification for Real-Time Audience Analysis System

Gender Classification for Real-Time Audience Analysis System Gender Classfcaton for Real-Tme Audence Analyss System Vladmr Khryashchev, Lev Shmaglt, Andrey Shemyakov, Anton Lebedev Yaroslavl State Unversty Yaroslavl, Russa vhr@yandex.ru, shmaglt_lev@yahoo.com, andrey.shemakov@gmal.com,

More information

Bayesian Network Based Causal Relationship Identification and Funding Success Prediction in P2P Lending

Bayesian Network Based Causal Relationship Identification and Funding Success Prediction in P2P Lending Proceedngs of 2012 4th Internatonal Conference on Machne Learnng and Computng IPCSIT vol. 25 (2012) (2012) IACSIT Press, Sngapore Bayesan Network Based Causal Relatonshp Identfcaton and Fundng Success

More information

A heuristic task deployment approach for load balancing

A heuristic task deployment approach for load balancing Xu Gaochao, Dong Yunmeng, Fu Xaodog, Dng Yan, Lu Peng, Zhao Ja Abstract A heurstc task deployment approach for load balancng Gaochao Xu, Yunmeng Dong, Xaodong Fu, Yan Dng, Peng Lu, Ja Zhao * College of

More information

Improved SVM in Cloud Computing Information Mining

Improved SVM in Cloud Computing Information Mining Internatonal Journal of Grd Dstrbuton Computng Vol.8, No.1 (015), pp.33-40 http://dx.do.org/10.1457/jgdc.015.8.1.04 Improved n Cloud Computng Informaton Mnng Lvshuhong (ZhengDe polytechnc college JangSu

More information

A Genetic Programming Based Stock Price Predictor together with Mean-Variance Based Sell/Buy Actions

A Genetic Programming Based Stock Price Predictor together with Mean-Variance Based Sell/Buy Actions Proceedngs of the World Congress on Engneerng 28 Vol II WCE 28, July 2-4, 28, London, U.K. A Genetc Programmng Based Stock Prce Predctor together wth Mean-Varance Based Sell/Buy Actons Ramn Rajaboun and

More information

ECE544NA Final Project: Robust Machine Learning Hardware via Classifier Ensemble

ECE544NA Final Project: Robust Machine Learning Hardware via Classifier Ensemble 1 ECE544NA Fnal Project: Robust Machne Learnng Hardware va Classfer Ensemble Sa Zhang, szhang12@llnos.edu Dept. of Electr. & Comput. Eng., Unv. of Illnos at Urbana-Champagn, Urbana, IL, USA Abstract In

More information

Master s Thesis. Configuring robust virtual wireless sensor networks for Internet of Things inspired by brain functional networks

Master s Thesis. Configuring robust virtual wireless sensor networks for Internet of Things inspired by brain functional networks Master s Thess Ttle Confgurng robust vrtual wreless sensor networks for Internet of Thngs nspred by bran functonal networks Supervsor Professor Masayuk Murata Author Shnya Toyonaga February 10th, 2014

More information

Design and Development of a Security Evaluation Platform Based on International Standards

Design and Development of a Security Evaluation Platform Based on International Standards Internatonal Journal of Informatcs Socety, VOL.5, NO.2 (203) 7-80 7 Desgn and Development of a Securty Evaluaton Platform Based on Internatonal Standards Yuj Takahash and Yoshm Teshgawara Graduate School

More information

Realistic Image Synthesis

Realistic Image Synthesis Realstc Image Synthess - Combned Samplng and Path Tracng - Phlpp Slusallek Karol Myszkowsk Vncent Pegoraro Overvew: Today Combned Samplng (Multple Importance Samplng) Renderng and Measurng Equaton Random

More information

ERP Software Selection Using The Rough Set And TPOSIS Methods

ERP Software Selection Using The Rough Set And TPOSIS Methods ERP Software Selecton Usng The Rough Set And TPOSIS Methods Under Fuzzy Envronment Informaton Management Department, Hunan Unversty of Fnance and Economcs, No. 139, Fengln 2nd Road, Changsha, 410205, Chna

More information

Software project management with GAs

Software project management with GAs Informaton Scences 177 (27) 238 241 www.elsever.com/locate/ns Software project management wth GAs Enrque Alba *, J. Francsco Chcano Unversty of Málaga, Grupo GISUM, Departamento de Lenguajes y Cencas de

More information

Mining Feature Importance: Applying Evolutionary Algorithms within a Web-based Educational System

Mining Feature Importance: Applying Evolutionary Algorithms within a Web-based Educational System Mnng Feature Importance: Applyng Evolutonary Algorthms wthn a Web-based Educatonal System Behrouz MINAEI-BIDGOLI 1, and Gerd KORTEMEYER 2, and Wllam F. PUNCH 1 1 Genetc Algorthms Research and Applcatons

More information

Calculating the high frequency transmission line parameters of power cables

Calculating the high frequency transmission line parameters of power cables < ' Calculatng the hgh frequency transmsson lne parameters of power cables Authors: Dr. John Dcknson, Laboratory Servces Manager, N 0 RW E B Communcatons Mr. Peter J. Ncholson, Project Assgnment Manager,

More information

An efficient constraint handling methodology for multi-objective evolutionary algorithms

An efficient constraint handling methodology for multi-objective evolutionary algorithms Rev. Fac. Ing. Unv. Antoqua N. 49. pp. 141-150. Septembre, 009 An effcent constrant handlng methodology for mult-objectve evolutonary algorthms Una metodología efcente para manejo de restrccones en algortmos

More information

Investigation of Modified Bee Colony Algorithm with Particle and Chaos Theory

Investigation of Modified Bee Colony Algorithm with Particle and Chaos Theory Internatonal Journal of Control and Automaton, pp. 311-3 http://dx.do.org/10.1457/jca.015.8..30 Investgaton of Modfed Bee Colony Algorthm wth Partcle and Chaos Theory Guo Cheng Shangluo College, Zhangye,

More information

A novel Method for Data Mining and Classification based on

A novel Method for Data Mining and Classification based on A novel Method for Data Mnng and Classfcaton based on Ensemble Learnng 1 1, Frst Author Nejang Normal Unversty;Schuan Nejang 641112,Chna, E-mal: lhan-gege@126.com Abstract Data mnng has been attached great

More information

LITERATURE REVIEW: VARIOUS PRIORITY BASED TASK SCHEDULING ALGORITHMS IN CLOUD COMPUTING

LITERATURE REVIEW: VARIOUS PRIORITY BASED TASK SCHEDULING ALGORITHMS IN CLOUD COMPUTING LITERATURE REVIEW: VARIOUS PRIORITY BASED TASK SCHEDULING ALGORITHMS IN CLOUD COMPUTING 1 MS. POOJA.P.VASANI, 2 MR. NISHANT.S. SANGHANI 1 M.Tech. [Software Systems] Student, Patel College of Scence and

More information

Multiple-Period Attribution: Residuals and Compounding

Multiple-Period Attribution: Residuals and Compounding Multple-Perod Attrbuton: Resduals and Compoundng Our revewer gave these authors full marks for dealng wth an ssue that performance measurers and vendors often regard as propretary nformaton. In 1994, Dens

More information

8.5 UNITARY AND HERMITIAN MATRICES. The conjugate transpose of a complex matrix A, denoted by A*, is given by

8.5 UNITARY AND HERMITIAN MATRICES. The conjugate transpose of a complex matrix A, denoted by A*, is given by 6 CHAPTER 8 COMPLEX VECTOR SPACES 5. Fnd the kernel of the lnear transformaton gven n Exercse 5. In Exercses 55 and 56, fnd the mage of v, for the ndcated composton, where and are gven by the followng

More information

Global Optimization Algorithms with Application to Non-Life Insurance

Global Optimization Algorithms with Application to Non-Life Insurance Global Optmzaton Algorthms wth Applcaton to Non-Lfe Insurance Problems Ralf Kellner Workng Paper Char for Insurance Economcs Fredrch-Alexander-Unversty of Erlangen-Nürnberg Verson: June 202 GLOBAL OPTIMIZATION

More information

Luby s Alg. for Maximal Independent Sets using Pairwise Independence

Luby s Alg. for Maximal Independent Sets using Pairwise Independence Lecture Notes for Randomzed Algorthms Luby s Alg. for Maxmal Independent Sets usng Parwse Independence Last Updated by Erc Vgoda on February, 006 8. Maxmal Independent Sets For a graph G = (V, E), an ndependent

More information

How To Know The Components Of Mean Squared Error Of Herarchcal Estmator S

How To Know The Components Of Mean Squared Error Of Herarchcal Estmator S S C H E D A E I N F O R M A T I C A E VOLUME 0 0 On Mean Squared Error of Herarchcal Estmator Stans law Brodowsk Faculty of Physcs, Astronomy, and Appled Computer Scence, Jagellonan Unversty, Reymonta

More information

J. Parallel Distrib. Comput.

J. Parallel Distrib. Comput. J. Parallel Dstrb. Comput. 71 (2011) 62 76 Contents lsts avalable at ScenceDrect J. Parallel Dstrb. Comput. journal homepage: www.elsever.com/locate/jpdc Optmzng server placement n dstrbuted systems n

More information

A Binary Particle Swarm Optimization Algorithm for Lot Sizing Problem

A Binary Particle Swarm Optimization Algorithm for Lot Sizing Problem Journal o Economc and Socal Research 5 (2), -2 A Bnary Partcle Swarm Optmzaton Algorthm or Lot Szng Problem M. Fath Taşgetren & Yun-Cha Lang Abstract. Ths paper presents a bnary partcle swarm optmzaton

More information

Joint Scheduling of Processing and Shuffle Phases in MapReduce Systems

Joint Scheduling of Processing and Shuffle Phases in MapReduce Systems Jont Schedulng of Processng and Shuffle Phases n MapReduce Systems Fangfe Chen, Mural Kodalam, T. V. Lakshman Department of Computer Scence and Engneerng, The Penn State Unversty Bell Laboratores, Alcatel-Lucent

More information

Abstract. Clustering ensembles have emerged as a powerful method for improving both the

Abstract. Clustering ensembles have emerged as a powerful method for improving both the Clusterng Ensembles: {topchyal, Models jan, of punch}@cse.msu.edu Consensus and Weak Parttons * Alexander Topchy, Anl K. Jan, and Wllam Punch Department of Computer Scence and Engneerng, Mchgan State Unversty

More information

Mooring Pattern Optimization using Genetic Algorithms

Mooring Pattern Optimization using Genetic Algorithms 6th World Congresses of Structural and Multdscplnary Optmzaton Ro de Janero, 30 May - 03 June 005, Brazl Moorng Pattern Optmzaton usng Genetc Algorthms Alonso J. Juvnao Carbono, Ivan F. M. Menezes Luz

More information

Fast Fuzzy Clustering of Web Page Collections

Fast Fuzzy Clustering of Web Page Collections Fast Fuzzy Clusterng of Web Page Collectons Chrstan Borgelt and Andreas Nürnberger Dept. of Knowledge Processng and Language Engneerng Otto-von-Guercke-Unversty of Magdeburg Unverstätsplatz, D-396 Magdeburg,

More information

Sensor placement for leak detection and location in water distribution networks

Sensor placement for leak detection and location in water distribution networks Sensor placement for leak detecton and locaton n water dstrbuton networks ABSTRACT R. Sarrate*, J. Blesa, F. Near, J. Quevedo Automatc Control Department, Unverstat Poltècnca de Catalunya, Rambla de Sant

More information

Period and Deadline Selection for Schedulability in Real-Time Systems

Period and Deadline Selection for Schedulability in Real-Time Systems Perod and Deadlne Selecton for Schedulablty n Real-Tme Systems Thdapat Chantem, Xaofeng Wang, M.D. Lemmon, and X. Sharon Hu Department of Computer Scence and Engneerng, Department of Electrcal Engneerng

More information

Intelligent Method for Cloud Task Scheduling Based on Particle Swarm Optimization Algorithm

Intelligent Method for Cloud Task Scheduling Based on Particle Swarm Optimization Algorithm Unversty of Nzwa, Oman December 9-11, 2014 Page 39 THE INTERNATIONAL ARAB CONFERENCE ON INFORMATION TECHNOLOGY (ACIT2014) Intellgent Method for Cloud Task Schedulng Based on Partcle Swarm Optmzaton Algorthm

More information

Web Object Indexing Using Domain Knowledge *

Web Object Indexing Using Domain Knowledge * Web Object Indexng Usng Doman Knowledge * Muyuan Wang Department of Automaton Tsnghua Unversty Bejng 100084, Chna (86-10)51774518 Zhwe L, Le Lu, We-Yng Ma Mcrosoft Research Asa Sgma Center, Hadan Dstrct

More information

A Programming Model for the Cloud Platform

A Programming Model for the Cloud Platform Internatonal Journal of Advanced Scence and Technology A Programmng Model for the Cloud Platform Xaodong Lu School of Computer Engneerng and Scence Shangha Unversty, Shangha 200072, Chna luxaodongxht@qq.com

More information

VoIP Playout Buffer Adjustment using Adaptive Estimation of Network Delays

VoIP Playout Buffer Adjustment using Adaptive Estimation of Network Delays VoIP Playout Buffer Adjustment usng Adaptve Estmaton of Network Delays Mroslaw Narbutt and Lam Murphy* Department of Computer Scence Unversty College Dubln, Belfeld, Dubln, IRELAND Abstract The poor qualty

More information

The Journal of Systems and Software

The Journal of Systems and Software The Journal of Systems and Software 82 (2009) 241 252 Contents lsts avalable at ScenceDrect The Journal of Systems and Software journal homepage: www. elsever. com/ locate/ jss A study of project selecton

More information

Hybrid-Learning Methods for Stock Index Modeling

Hybrid-Learning Methods for Stock Index Modeling Hybrd-Learnng Methods for Stock Index Modelng 63 Chapter IV Hybrd-Learnng Methods for Stock Index Modelng Yuehu Chen, Jnan Unversty, Chna Ajth Abraham, Chung-Ang Unversty, Republc of Korea Abstract The

More information

Vehicle Detection and Tracking in Video from Moving Airborne Platform

Vehicle Detection and Tracking in Video from Moving Airborne Platform Journal of Computatonal Informaton Systems 10: 12 (2014) 4965 4972 Avalable at http://www.jofcs.com Vehcle Detecton and Trackng n Vdeo from Movng Arborne Platform Lye ZHANG 1,2,, Hua WANG 3, L LI 2 1 School

More information

A Novel Methodology of Working Capital Management for Large. Public Constructions by Using Fuzzy S-curve Regression

A Novel Methodology of Working Capital Management for Large. Public Constructions by Using Fuzzy S-curve Regression Novel Methodology of Workng Captal Management for Large Publc Constructons by Usng Fuzzy S-curve Regresson Cheng-Wu Chen, Morrs H. L. Wang and Tng-Ya Hseh Department of Cvl Engneerng, Natonal Central Unversty,

More information

RESEARCH ON DUAL-SHAKER SINE VIBRATION CONTROL. Yaoqi FENG 1, Hanping QIU 1. China Academy of Space Technology (CAST) yaoqi.feng@yahoo.

RESEARCH ON DUAL-SHAKER SINE VIBRATION CONTROL. Yaoqi FENG 1, Hanping QIU 1. China Academy of Space Technology (CAST) yaoqi.feng@yahoo. ICSV4 Carns Australa 9- July, 007 RESEARCH ON DUAL-SHAKER SINE VIBRATION CONTROL Yaoq FENG, Hanpng QIU Dynamc Test Laboratory, BISEE Chna Academy of Space Technology (CAST) yaoq.feng@yahoo.com Abstract

More information

Vehicle Routing Problem with Time Windows for Reducing Fuel Consumption

Vehicle Routing Problem with Time Windows for Reducing Fuel Consumption 3020 JOURNAL OF COMPUTERS, VOL. 7, NO. 12, DECEMBER 2012 Vehcle Routng Problem wth Tme Wndows for Reducng Fuel Consumpton Jn L School of Computer and Informaton Engneerng, Zhejang Gongshang Unversty, Hangzhou,

More information

Parallel Numerical Simulation of Visual Neurons for Analysis of Optical Illusion

Parallel Numerical Simulation of Visual Neurons for Analysis of Optical Illusion 212 Thrd Internatonal Conference on Networkng and Computng Parallel Numercal Smulaton of Vsual Neurons for Analyss of Optcal Illuson Akra Egashra, Shunj Satoh, Hdetsugu Ire and Tsutomu Yoshnaga Graduate

More information

Data Broadcast on a Multi-System Heterogeneous Overlayed Wireless Network *

Data Broadcast on a Multi-System Heterogeneous Overlayed Wireless Network * JOURNAL OF INFORMATION SCIENCE AND ENGINEERING 24, 819-840 (2008) Data Broadcast on a Mult-System Heterogeneous Overlayed Wreless Network * Department of Computer Scence Natonal Chao Tung Unversty Hsnchu,

More information

SOLVING CARDINALITY CONSTRAINED PORTFOLIO OPTIMIZATION PROBLEM BY BINARY PARTICLE SWARM OPTIMIZATION ALGORITHM

SOLVING CARDINALITY CONSTRAINED PORTFOLIO OPTIMIZATION PROBLEM BY BINARY PARTICLE SWARM OPTIMIZATION ALGORITHM SOLVIG CARDIALITY COSTRAIED PORTFOLIO OPTIMIZATIO PROBLEM BY BIARY PARTICLE SWARM OPTIMIZATIO ALGORITHM Aleš Kresta Klíčová slova: optmalzace portfola, bnární algortmus rojení částc Key words: portfolo

More information

Project Networks With Mixed-Time Constraints

Project Networks With Mixed-Time Constraints Project Networs Wth Mxed-Tme Constrants L Caccetta and B Wattananon Western Australan Centre of Excellence n Industral Optmsaton (WACEIO) Curtn Unversty of Technology GPO Box U1987 Perth Western Australa

More information

A Prefix Code Matching Parallel Load-Balancing Method for Solution-Adaptive Unstructured Finite Element Graphs on Distributed Memory Multicomputers

A Prefix Code Matching Parallel Load-Balancing Method for Solution-Adaptive Unstructured Finite Element Graphs on Distributed Memory Multicomputers Ž. The Journal of Supercomputng, 15, 25 49 2000 2000 Kluwer Academc Publshers. Manufactured n The Netherlands. A Prefx Code Matchng Parallel Load-Balancng Method for Soluton-Adaptve Unstructured Fnte Element

More information

A DYNAMIC CRASHING METHOD FOR PROJECT MANAGEMENT USING SIMULATION-BASED OPTIMIZATION. Michael E. Kuhl Radhamés A. Tolentino-Peña

A DYNAMIC CRASHING METHOD FOR PROJECT MANAGEMENT USING SIMULATION-BASED OPTIMIZATION. Michael E. Kuhl Radhamés A. Tolentino-Peña Proceedngs of the 2008 Wnter Smulaton Conference S. J. Mason, R. R. Hll, L. Mönch, O. Rose, T. Jefferson, J. W. Fowler eds. A DYNAMIC CRASHING METHOD FOR PROJECT MANAGEMENT USING SIMULATION-BASED OPTIMIZATION

More information

A Binary Quantum-behaved Particle Swarm Optimization Algorithm with Cooperative Approach

A Binary Quantum-behaved Particle Swarm Optimization Algorithm with Cooperative Approach IJCSI Internatonal Journal of Computer Scence Issues, Vol., Issue, No, January 3 ISSN (Prnt): 694-784 ISSN (Onlne): 694-84 www.ijcsi.org A Bnary Quantum-behave Partcle Swarm Optmzaton Algorthm wth Cooperatve

More information

CHOLESTEROL REFERENCE METHOD LABORATORY NETWORK. Sample Stability Protocol

CHOLESTEROL REFERENCE METHOD LABORATORY NETWORK. Sample Stability Protocol CHOLESTEROL REFERENCE METHOD LABORATORY NETWORK Sample Stablty Protocol Background The Cholesterol Reference Method Laboratory Network (CRMLN) developed certfcaton protocols for total cholesterol, HDL

More information

Comparison of Weighted Sum Fitness Functions for PSO Optimization of Wideband Medium-gain Antennas

Comparison of Weighted Sum Fitness Functions for PSO Optimization of Wideband Medium-gain Antennas 54 ZHOGKU MA, G. A. E. VAEBOSCH, COMPARISO OF WEIGHTE SUM FITESS FUCTIOS FOR PSO Comparson of Weghted Sum Ftness Functons for PSO Optmzaton of Wdeband Medum-gan Antennas Zhongkun MA, Guy A. E. VAEBOSCH

More information

Optimal Choice of Random Variables in D-ITG Traffic Generating Tool using Evolutionary Algorithms

Optimal Choice of Random Variables in D-ITG Traffic Generating Tool using Evolutionary Algorithms Optmal Choce of Random Varables n D-ITG Traffc Generatng Tool usng Evolutonary Algorthms M. R. Mosav* (C.A.), F. Farab* and S. Karam* Abstract: Impressve development of computer networks has been requred

More information

BERNSTEIN POLYNOMIALS

BERNSTEIN POLYNOMIALS On-Lne Geometrc Modelng Notes BERNSTEIN POLYNOMIALS Kenneth I. Joy Vsualzaton and Graphcs Research Group Department of Computer Scence Unversty of Calforna, Davs Overvew Polynomals are ncredbly useful

More information

A Data Mining-Based OLAP Aggregation of. Complex Data: Application on XML Documents

A Data Mining-Based OLAP Aggregation of. Complex Data: Application on XML Documents 1 Runnng head: A DATA MINING-BASED OLAP AGGREGATION A Data Mnng-Based OLAP Aggregaton of Complex Data: Applcaton on XML Documents Radh Ben Messaoud, Omar Boussad, Sabne Loudcher Rabaséda {rbenmessaoud

More information

Single and multiple stage classifiers implementing logistic discrimination

Single and multiple stage classifiers implementing logistic discrimination Sngle and multple stage classfers mplementng logstc dscrmnaton Hélo Radke Bttencourt 1 Dens Alter de Olvera Moraes 2 Vctor Haertel 2 1 Pontfíca Unversdade Católca do Ro Grande do Sul - PUCRS Av. Ipranga,

More information

Ant Colony Optimization for Economic Generator Scheduling and Load Dispatch

Ant Colony Optimization for Economic Generator Scheduling and Load Dispatch Proceedngs of the th WSEAS Int. Conf. on EVOLUTIONARY COMPUTING, Lsbon, Portugal, June 1-18, 5 (pp17-175) Ant Colony Optmzaton for Economc Generator Schedulng and Load Dspatch K. S. Swarup Abstract Feasblty

More information

Characterization of Assembly. Variation Analysis Methods. A Thesis. Presented to the. Department of Mechanical Engineering. Brigham Young University

Characterization of Assembly. Variation Analysis Methods. A Thesis. Presented to the. Department of Mechanical Engineering. Brigham Young University Characterzaton of Assembly Varaton Analyss Methods A Thess Presented to the Department of Mechancal Engneerng Brgham Young Unversty In Partal Fulfllment of the Requrements for the Degree Master of Scence

More information

A Fast Incremental Spectral Clustering for Large Data Sets

A Fast Incremental Spectral Clustering for Large Data Sets 2011 12th Internatonal Conference on Parallel and Dstrbuted Computng, Applcatons and Technologes A Fast Incremental Spectral Clusterng for Large Data Sets Tengteng Kong 1,YeTan 1, Hong Shen 1,2 1 School

More information

Study on Model of Risks Assessment of Standard Operation in Rural Power Network

Study on Model of Risks Assessment of Standard Operation in Rural Power Network Study on Model of Rsks Assessment of Standard Operaton n Rural Power Network Qngj L 1, Tao Yang 2 1 Qngj L, College of Informaton and Electrcal Engneerng, Shenyang Agrculture Unversty, Shenyang 110866,

More information

Brigid Mullany, Ph.D University of North Carolina, Charlotte

Brigid Mullany, Ph.D University of North Carolina, Charlotte Evaluaton And Comparson Of The Dfferent Standards Used To Defne The Postonal Accuracy And Repeatablty Of Numercally Controlled Machnng Center Axes Brgd Mullany, Ph.D Unversty of North Carolna, Charlotte

More information

What is Candidate Sampling

What is Candidate Sampling What s Canddate Samplng Say we have a multclass or mult label problem where each tranng example ( x, T ) conssts of a context x a small (mult)set of target classes T out of a large unverse L of possble

More information

The OC Curve of Attribute Acceptance Plans

The OC Curve of Attribute Acceptance Plans The OC Curve of Attrbute Acceptance Plans The Operatng Characterstc (OC) curve descrbes the probablty of acceptng a lot as a functon of the lot s qualty. Fgure 1 shows a typcal OC Curve. 10 8 6 4 1 3 4

More information

Recurrence. 1 Definitions and main statements

Recurrence. 1 Definitions and main statements Recurrence 1 Defntons and man statements Let X n, n = 0, 1, 2,... be a MC wth the state space S = (1, 2,...), transton probabltes p j = P {X n+1 = j X n = }, and the transton matrx P = (p j ),j S def.

More information

Maintenance Scheduling by using the Bi-Criterion Algorithm of Preferential Anti-Pheromone

Maintenance Scheduling by using the Bi-Criterion Algorithm of Preferential Anti-Pheromone Leonardo ournal of Scences ISSN 583-0233 Issue 2, anuary-une 2008 p. 43-64 Mantenance Schedulng by usng the B-Crteron Algorthm of Preferental Ant-Pheromone Trantafyllos MYTAKIDIS and Arstds VLACHOS Department

More information

universitat Autónoma' de Barcelona

universitat Autónoma' de Barcelona unverstat Autónoma' de Barcelona A new dstrbuted dffuson algorthm for dynamc load-balancng n parallel systems Departament d'informàtca Untat d'arqutectura d'ordnadors Sstemes Operatus A thess submtted

More information

FREQUENCY OF OCCURRENCE OF CERTAIN CHEMICAL CLASSES OF GSR FROM VARIOUS AMMUNITION TYPES

FREQUENCY OF OCCURRENCE OF CERTAIN CHEMICAL CLASSES OF GSR FROM VARIOUS AMMUNITION TYPES FREQUENCY OF OCCURRENCE OF CERTAIN CHEMICAL CLASSES OF GSR FROM VARIOUS AMMUNITION TYPES Zuzanna BRO EK-MUCHA, Grzegorz ZADORA, 2 Insttute of Forensc Research, Cracow, Poland 2 Faculty of Chemstry, Jagellonan

More information

Cloud-based Social Application Deployment using Local Processing and Global Distribution

Cloud-based Social Application Deployment using Local Processing and Global Distribution Cloud-based Socal Applcaton Deployment usng Local Processng and Global Dstrbuton Zh Wang *, Baochun L, Lfeng Sun *, and Shqang Yang * * Bejng Key Laboratory of Networked Multmeda Department of Computer

More information

An Evaluation of the Extended Logistic, Simple Logistic, and Gompertz Models for Forecasting Short Lifecycle Products and Services

An Evaluation of the Extended Logistic, Simple Logistic, and Gompertz Models for Forecasting Short Lifecycle Products and Services An Evaluaton of the Extended Logstc, Smple Logstc, and Gompertz Models for Forecastng Short Lfecycle Products and Servces Charles V. Trappey a,1, Hsn-yng Wu b a Professor (Management Scence), Natonal Chao

More information

Exhaustive Regression. An Exploration of Regression-Based Data Mining Techniques Using Super Computation

Exhaustive Regression. An Exploration of Regression-Based Data Mining Techniques Using Super Computation Exhaustve Regresson An Exploraton of Regresson-Based Data Mnng Technques Usng Super Computaton Antony Daves, Ph.D. Assocate Professor of Economcs Duquesne Unversty Pttsburgh, PA 58 Research Fellow The

More information