A New Hybrid Method Based on Improved Particle Swarm Optimization, Ant Colony Algorithm and HMM for Web Information Extraction

Size: px
Start display at page:

Download "A New Hybrid Method Based on Improved Particle Swarm Optimization, Ant Colony Algorithm and HMM for Web Information Extraction"

Transcription

1 A New Hybrd Method Based on Improved Partcle Swarm Optmzaton, Ant Colony Algorthm and HMM for Web Informaton Extracton Rong LI, Hong-bn WANG Department of Computer, Xnzhou Teachers Unversty, Xnzhou, Chna Abstract In order to further enhance the accuracy of Web nformaton extracton, and overcome the shortcomngs of the Hdden Markov Model (HMM) and ts hybrd method n parameter optmzaton, a novel Web extracton algorthm based on a combned and mproved partcle swarm optmzaton, ant colony algorthm (IPSO-ACA) and HMM s presented. Frst, an HMM for nformaton extracton s bult. Second, an mproved hybrd ntellgent algorthm combnng PSO wth ACA s proposed. In the new algorthm, nertal weghts of partcle swarm optmzaton and parameters of ant colony algorthm such as stmulatng factor, volatlzaton coeffcents and pheromones are all mproved adaptvely, and then the ftness functon values of partcles hstory optmal solutons are used to adjust the ntal pheromone dstrbuton of the ant colony algorthm. Thrd, the hybrd ntellgent algorthm s adopted for the approxmate global optmal soluton and then Baum-Welch algorthm (BW) s adopted for the local modfcaton, whch not only solves the BW dependency on ntal values and the trapped local optmum problem, but also makes full use of the global search ablty of the hybrd ntellgent algorthm and local development ablty of BW. Fnally, the Vterb algorthm s used to decode the HMM model. Compared wth exstng HMM optmzaton methods, the comprehensve Fβ=1 value s averagely ncreased by 7.3%, whch shows that the mproved algorthm can effectvely enhance optmzaton performance and extracton accuracy. Keywords - nformaton extracton, Hdden Markov Model, partcle swarm optmzaton, ant colony algorthm I. INTRODUCTION Wth the development of Internet technology, Web resources have revealed the ncrease trend n massve and unstructured nformaton. How to recognze those nterestng data for users from unstructured or sem-structured Web nformaton and turn them nto a more structured and clearer semantc format, whch s the technology problem of Web nformaton extracton[1]. A large number of experts and scholars appled statstcal machne learnng methods to ths feld. Typcal statstcal methods manly nclude Hdden Markov Model (HMM)[2-3] and ts hybrd methods[4-8]. For nstance, Zhang et al.[4] Combned bnary HMM and SVM to realze the metadata extracton. Ln et al.[5] proposed text nformaton extracton method based on maxmum entropy and HMM, whch used the weghtng sum of observaton text feature to adjust HMM transton probablty. Xao et al.[6] employed genetc algorthm(ga) to optmze HMM parameters and obtaned the extracton effect superor to tradtonal HMM, but the approach stll reflected the precocous shortcomng of GA. Zou et al.[7] proposed Web nformaton extracton based on smulated annealng(sa) and HMM, but the method dd not consder HMM context features. Wang et al.[8] presented a web extracton algorthm usng mproved PSO and HMM, ts mprovement emboded n nerta weght and the mutaton of part partcles. By Analyss of these HMM lteratures, there s stll lots of room for mprovement n parameter optmzaton and extracton performance. In Inspred by ths, ths paper proposes a self-adaptve hybrd ntellgent optmzaton HMM algorthm for Web ctaton extracton. After constructng an HMM, a selfadaptve hybrd ntellgent optmzaton algorthm based on mproved PSO and ACA(or IPSAA for short) s put forward. The new algorthm realzes the dynamc self-adaptve adjustment n parameters, such as nertal weght of PSO and stmulatng factor, volatlzaton coeffcents and pheromone of ACA, then the ftness functon values of partcles hstory optmal solutons are used to adjust the ntal pheromone dstrbuton of ant colony algorthm, whch s the pont cut of juncture between PSO and ACA. And then ths paper maps the approxmate global optmal soluton, found out by IPSAA, as the ntal model of BW algorthm and adopts BW to contnue to modfy locally parameters. Fnally, the mproved model uses Vterb algorthm to decode for the optmal state sequences. Expermental results ndcate that the IPSAA-HMM algorthm greatly mproves the accuracy of Web nformaton extracton, whch proves the feasblty and effectveness of the IPSAA-HMM algorthm. II. WEB INFORMATION EXTRACTION BASED ON HMM An HMM may be vewed as a fve-tuple (S,O,Π,A,B),where 1S s a state set contanng N states, denoted as S={S 1,S 2,,S N };2O s a symbol set ncludng M output symbols, denoted as O={O 1,O 2, O M};3Π s the ntal state probablty matrx, denoted as N ={ }, =P(q 1=S ), 1 N,0 1, 1 1 ;4A s the state transton probablty matrx; 5 B s symbol output probablty matrx, denoted as M B={b j(o k )},b j(o k )=P(o t=v k q t=s j ), 0 b j(o k ) 1, b j(o k ) 1,1 j N,1k M. k1 DOI /IJSSST.a ISSN: x onlne, prnt

2 When HMM s appled to nformaton extracton, model observaton layer s the text sequence to be observed, hdden layer s state sequence composed of state domans such as <Author>, <Ttle> and <Journal>. Extracton process may be descrbed as follows: gven the HMM model λ=(π,a,b )and the observe text sequence O=(O 1,O 2, O T ), ntalze HMM parameters randomly, and use BW algorthm for HMM tranng so as to buld HMM, fnally adopt Vterb algorthm to fnd out the state doman sequence q*=(q 1,q 2,,q T} wth the maxmum probablty P(q O). III. HMM TRAINING ALGORITHM BASED ON IPSAA A. Partcle Swarm Algorthm and Basc Ant Colony Algorthm Partcle swarm optmzaton(pso) algorthm s a global optmzaton algorthm smulatng the movement behavor of brd swarm[9-11]. The soluton of each problem may be seen as a partcle n search pace. Partcles update ther velocty and poston by trackng the ndvdual optmal soluton x pbest and the global optmal soluton x gbest, the updatng equatons are as follows: v(t) v(t 1) cr(t)(x 11 pbest x(t)) c r(t)(x x(t)) (1) 22 gbest x(t) x (t 1) v(t) (2) Where, x (t) and v (t) represent the locaton and velocty of the th partcle n the t th generaton, ω represents nonnegatve nerta weght, c 1 and c 2 are non-negatve learnng factors, and r 1 r 2 are random numbers n [0,1]. Ant colony algorthm s a bonc evolutonary heurstc algorthm, whch s proposed by Dorgo. Through pheromone-nduced effect, ndvdual ants make the later ants choose the shorter path wth stronger pheromone and the algorthm gradually converges to the global optmal soluton. Frst, m ants are randomly placed n n nodes, the probablty of whch the k th ant n note selects the next note j s as follows: [( j( t)] [ j ] k p () t [( ( )] [ ], j l t l j alloedk (3) luk 0, else In Eq.(3), j () t represents the amount of pheromones between and j notes at tme t. j() t ndcates heurstc functon. Parametersα vs β, pheromone heurstc factor vs desred heurstc factor, are used for determnng the sgnfcance between amount of pheromones and dstance nter-node. allowed k ndcates the next note set that ant k s allowed to select. As tme goes on, pheromones left before gradually dsappear. After n moment, ants complete 1 cycle, and the amount of pheromones on each path should be adjusted accordng to Eq.(4). j( t n) j ( t) j, ( 01, ) m k j j k 1 (4) k Q/ lk, j Lk j 0, else k Where j () t ndcates the amount of pheromones left by the k th ant n ths teraton. j ndcates the pheromone ncrement between notes and j n the cycle. ρ s the pheromone evaporaton coeffcent and Q s a constant. L k and l k respectvely ndcate the k th ant s traveled path and length n ths teraton. B. Improvement Strategy of PSO In ths artcle, the mprovement of pso s manly amed at adaptve nerta weght adjustment. The value of nerta weght ω has mportant nfluence on the PSO optmzaton search. Generally, n order to obtan better algorthm performance, usually n the search early stage ω should has a greater value to ensure the partcle swarms strong global search ablty wthn a larger search space and avod premature. And as teratons ncrease, ω should has a smaller value to ensure the partcle swarms local search ablty wthn a smaller search space and enhance the convergence precson. Therefore, the approprate control of nerta weght n the teratve process can balance the global search and local search of algorthm, thereby gettng good enough soluton on average wth less teraton. Tradtonal self-adaptve method makes ω lnearly decrease wth the ncrease of teratons, whch mproves the performance of algorthm, but there are stll some shortcomngs. On the one hand, ths knd of PSO algorthm cant effectvely reflect the complcated non-lnear behavor n the partcle swarms actual search process, so the convergence speed and convergence precson s stll not deal. On the other hand, the slope of whch ω lnearly decreases s stll problem-dependency, there s no unversal optmal change slope for all optmzaton problem. By the prevous analyss, the change process of ω s dynamc and non-lnear, therefore, so ths paper adopts the non-lnear functon to descrbe the dynamc change rule of ω n the teraton process. The ω value of each teratve step s determned by the followng exponental functon formula: ternow n wter ( ) wnt exp( ( ) ) (5) termax DOI /IJSSST.a ISSN: x onlne, prnt

3 In Eq (5), n s a control power exponent of non-lnear change rule, partcularly when n=2, (5) s often referred to as probablty curve functon. Fgure (1) shows the ω teraton change curve wth dfferent n value. ω value n=1.25 n=1 n=2 n=1.75 n=1.5 n= ter now /ter max Fgure 1. ω Iteraton Change Curve wth Dfferent n Value. A shown n Fg.(1)., for the gven ntal value ω nt and the control power exponent n, the non-lnear change rule of ω wth teratons can be unquely determned. And the greater the n value s, the longer the global search duraton of partcle swarm s, whle the smaller the n value s, the longer the local search duraton of partcle swarm s. By d 2 usng d-dmenson sphercal functon f( x) x (d=6,x 1 [-5.12,5.12]) to valdate parameter values, the result shows that when ω nt value s set n [0.2,0.5], and n value s set n [0.5,2], the algorthm has excellent performance. C. Improvement Strategy of ACA Seekng a balance between "exploraton" and "explotaton" s one of the key ssues n the study of ant colony algorthm[12]. In order to fnd a balance between obtanng a new path and usng pror knowledge, two aspects should be consdered for the mprovements of ACA. On the one hand, the search space of ACA can be made as large as possble to search the soluton regon of possble optmal soluton. On the other sde, the current effectve nformaton wthn ant colony should be taken full advantage of so that the searchng emphass of ant colony algorthm focuses on ndvdual ntervals wth hgher ftness value, and thus the algorthm may converge to the global optmal soluton wth the greater probablty. The convergence speed of ant colony algorthm should be mproved as far as possble under the premse of fndng the global optmal soluton. In ths paper, an adaptve strategy s adopted to resolve the man contradcton between performance and convergence rate. 1) Adaptve adjustment of pheromone heurstc factorαand desred heurstc factor β When α=0, only path pheromone works, the algorthm s equvalent to the shortest path searchng, whch s the tradtonal greedy algorthm. And when β=0, the heurstc functon of path pheromone s 0, the algorthm s equvalent to the blnd random search, whch s purely heurstc algorthm wth postve feedback. At frst the ants do not understand the stuaton of the lnk, the pheromone on the lnk has lttle effect on wayfndng ants. Along wth the ncreasng of teraton tmes, the pheromone on the lnk s more and more mportant to wayfndng ants, In the end, the probablty of whch the wnner lnk s selected s larger and larger, thus the convergence speed of the wnner lnk s faster and faster, and fnally the optmal path s found. So the α and the β value may be adaptvely adjusted accordng to Eq.(6), such ncentve mechansm can speed up convergence and mprove search qualty. 54e 1 4e NC NC 2) Adaptve mprovement of evaporaton coeffcent ρ When the problem scale s large, the pheromone evaporaton coeffcent ρ makes the amount of nformaton of soluton, whch has never been searched, be reduced to close to zero, and reduces the global search ablty of algorthm. If ρ s too large, then when the amount of nformaton of the soluton ncreases, the selected lkelhood of those prevous solutons s too large, the global search ablty of algorthm wll declne. If ρ s too small, the convergence speed of algorthm wll be too slow. Based on the comprehensve consderaton for the global search ablty and convergence speed, ρ may be changed to threshold functon, namely when the algorthm optmum value does not sgnfcantly mprove wthn N cycles, ρ s updated accordng to the followng functon. mn (6) (), t () t ( t n) mn else (7) Where, the ntal value of ρ s 1. The mnmum value of ρ, mn, can prevent too small a ρ value from reducng the convergence speed of algorthm. γ ndcates volatle constrant coeffcent and γ (0,1]. In order to reasonably select ρ value, γ s expressed as a gradual process, whch can makeγ value dynamcally reduce wth ncreasng number of teratons. Its functon s as follows: ( ter) 2 ter ter 1 e max 0 In Eq.(8), 0 s an ntal maxmum of.φ s a postve coeffcent of adjustng the changng speed of. Iter s the teraton steps or search tmes. ter max s total number of teratons. 3) Adaptve mprovement of pheromone The exstence of pheromone evaporaton coeffcent makes the amount of pheromone on those paths, whch have never been searched, close to zero, thereby reducng the search ablty on these paths. On the contrary, when (8) DOI /IJSSST.a ISSN: x onlne, prnt

4 pheromone on a path s larger, the amount of nformaton on these paths ncreases, the opportunty of agan selectng those paths wll become larger, whch also affects the global search ablty of algorthm. In alluson to the problem, the pheromone value may be changed and updated accordng to formula (9). 1 ( m) ( 1 ) j ( t) j, f 1 ( m) ( 1 ) j ( t) j, f max j ( t 1) (9) max Where ψ(m) s a functon proportonal to the number of convergence m, the more the number of teratons s, the greater the value of ψ(m) s, such as ψ(m)=m/ct, where ct s a constant and m ndcates the number of contnuous convergence. In ths way, the algorthm dynamcally updates pheromone accordng to the dstrbuton of solutons, consequently dynamcally adjusts the pheromone ntensty on each path, so that the ants nether too concentrated nor too decentralzed, and thus avodng premature and local convergence and mprovng the global search capablty. D. HMM Optmzaton Tranng of Improved Partcle Swarm-Ant colony Algorthm (IPSAA) Partcle swarm optmzaton algorthm has better searchng ablty and faster convergence speed, but t has no advantage n the combnatoral optmzaton problems. Ant colony algorthm can make up for ths shortcomng, but t has the shortcomngs such as blndness and slow search speed n the ntal search. Ths paper combnes them and gves play to the complementary advantages to optmze HMM parameter λ = (π, A, B). 1) Connecton of PSO and ACA In the IPSAA algorthm, the ntal poston of ant corresponds to the optmal poston of each partcle n PSO. The ftness functon value of each partcle s hstory optmal soluton s used to adjust the ntal dstrbuton of pheromone n ant colony algorthm ACA, and the ACA ntal pheromone formula s shown as follows. f( x ) ka (10) s mn In Eq.(10), ndcates the mnmum pheromone mn constant. x ndcates the ant poston correspondng to the optmal partcle poston, and f(x ) s ts ftness functon. ( ) ka f x s the pheromone value converted from the PSO result,k s a constant greater than zero, 0<a 1. And thus, the greater f(x ) s, the more pheromones s here. 2) Coarse search of PSO for HMM parameters rough optmzaton In the frst phase, partcle swarm optmzaton s used to optmze HMM parameters. A partcle corresponds to an HMM, and the elements of partcles poston vector X s the lnear arrangement of HMM parameter λ = (π, A, B). Therefore, the dmenson of partcles search space s the sum of number of A, B andπelements, a total of N+N*N+N*M dmensons, and each partcle s a real coded strng of N+N*N+N*M dmensons. The optmal soluton wth the maxmum ftness, X best, can be obtaned by PSO coarse search. In ths algorthm, In order to make all the sequence wth the maxmum probablty, to make the model better explan the observed sequence, the logarthmc mean of probablty of the observed sequence s used to measure the qualty of the model. The ftness functon s as follows: L 1 k L k1 f( x ) f( ) ln( po ( )) (11) Where, λ s a composte HMM correspondng to the th partcle. L s the number of sequence observatons, O k ndcates one of the observaton sequences, whose length s T. The probablty value p(o λ ) can be calculated by forward-backward algorthm of HMM. Takng logarthm of the probablty s to avod an underflow probablty multplcaton of. Defnton (11) of ftness functon can also be appled to the subsequent colony algorthm. Constrants of HMM s probablty parameters λ s nonnegatve, [0,1], and that sum of probabltes should be one. Those partcles of dssatsfyng the constrants n each generaton need to be normalzed., that s, negatve probablty should be set to 0, all the probabltes a j should N be replaced by a a / a to meet the requrement that sum j j j j1 of all probabltes should be one. The parameters of Vectorπand matrx B are normalzed n the same way. 3) Fne search of ACA for HMM parameters elaborate optmzaton In the second stage, ant colony algorthm s used to search elaborately for the further optmal soluton of HMM parameters. Fnally, Baum-Welch algorthm s adopted for local modfcaton, and the fnal HMM optmzaton parameter λ s obtaned. In Ant colony algorthm, a contnuous search space Ω, on behalf of a set of all HMM parameters A, B, π, s frstly establshed. The dmenson of search space s the sum of dmensons of A, B and π, that s N*M+N*N+ N dmensons. It can be expressed as x=[π 1,,π N, a 11,,a NN,b 11,,b NM ] T,, whose parameters are the same as the HMM s parameters. X can also be smply expressed as: x= (x 1,x 2,,x n ), 0 x 1, =1,2,,n A pont n the space represents a soluton, once the correspondng x s determned, the values of HMM s A, B and π can be determned. Accordng to the forward backward algorthm, the value of correspondng P(O λ) can L be calculated. Let ( ) ( ) 1 k f x f ln( po ( )), the algorthm L k1 uses t to search the correspondng pont, so that the value of f(x) s the maxmum, then the correspondng λ value of HMM can be determned. In ant colony algorthm, the search s manly dvded nto two operatons: searchng soluton and updatng pheromone. The IPSAA algorthm n the paper was mproved and extended manly for the two operatons. Ants search globally accordng to regonal probablty selecton rules, and search locally and randomly wthn the radus of δ at the same tme, DOI /IJSSST.a ISSN: x onlne, prnt

5 through whch ants move to fnd the optmal feasble soluton. Once an ant fnds a better soluton, t modfes the relevant pheromone concentraton to attract other ants to further search. a) Search operaton IPSAA algorthm assumes the exstence of the ant colony Q consstng of m ants, ts task s to fnd the current optmal pont X best,whose functon F(X best ) s the largest, n the soluton space composed of HMM parameters. By Eq.(9), the pheromone content s assgned to the ntal value n the regon ntalzaton phase, and each regon s represented by ts center pont poston x. Ants traverse these areas nstead of searchng the entre space, assumng that the set of the regonal mdpont s X R. In each round of search, m ants are allocated to each regon for the optmal soluton search. Ants choose the area accordng to the probablty decson rule whch s a functon of the local avalable pheromone and heurstc nformaton. The mprovement s as follows: (1)Decson Rules for Regonal probablty f px ( X ) [( ( x)] [ ( x)] [( ( x)] [ ( x)] (12) (( x ) Where ndcates the pheromone content of regonal center. ( x ) ndcates the heurstc nformaton, whch s f(x ) correspondng to the regonal center pont x. It can be calculated by Eq.(12), α, β are all postve parameters whch determne the role of pheromone and heurstc nformaton on the role of the selecton probablty. The larger α value s, the more the algorthm s nclned to the development of known search experence. Whereas the larger β value s, the stronger exploratory ablty the algorthm has. Ants choose the area accordng to the probablty decson rule and ants frst are located on the center pont of the selected area. Based on the dea of API algorthm[13], the area center pont s regarded as an ant nest, several ponts n the regon are selected as huntng spots. To ensure that the generated huntng ponts also satsfy the constrants of HMM learnng problems, the paper ntroduces a feasble soluton generaton rule to generate the relevant huntng spots and other search ponts. (2)Generaton Rule for Feasble Solutons Let nput pont be ( 1, 2,..., x x x x n ), where x s the vector element of the pont, s the current dmenson, and vbraton varable δ [0,r]. For x when [1,N], the p algorthm selects C N feasble ponts and makes each p x x,then selects C feasble ponts and makes each N x x,p [0,N/2]. For x when [N+1,N+N], the algorthm does the same thng. Then the algorthm postpones the nterval n bts, and then does the same thng, and so on, untl N*N+N. For x when [N*N+N+1, p N*N+N+N], the algorthm selects C M feasble ponts and p makes each x x,then selects C feasble ponts and M makes each x x,p [0,M/2]. The algorthm postpones the nterval n bts and does the same thng, and so on, untl N*N+N*M+N+1. And then the algorthm wll analyze whether each vector value of the newly generated pont x meets the condtons x r x x r, those ponts whch don t meet the condtons wll be abandoned. Based on Rule 2, a pont set Ω composed of N / 2 N / 2 P P P P NM ( C N C N ) ( CM CM ) P 0 P 0 dots can be generated. The δ value s enlarged or reduced n a certan vbraton order to make the selecton of the pont be unformly dstrbuted throughout the regon. The new algorthm randomly selects p ponts as huntng spots from Ω. b) Pheromone update operaton At the begnnng of each round of search, ants frst select the regon n accordance wth the pheromone dstrbuton of each regon. Regonal pheromone content s equal to the sum of pheromone contents of each huntng ponts n the regon. Ant departures from the nest, randomly selects a huntng pont to start the search. In the search process, f the s b operaton of x x happens, whch means the ant fnds the better huntng pont than the current one at the end of search near the huntng pont, the algorthm replaces the orgnal huntng pont wth the current pont and ncreases the pheromones correspondng to the current huntng spot. The update formula of pheromone ncrement s as shown n Eq.(13). j j Q (13) In the IPSAA algorthm, ants select area to search based on the regon pheromone content, and n the area dfferent ants also exchange nformaton. The nternal nformaton nteracton gudes ants to search near the huntng pont of better ftness functon. 4) HMM Tranng algorthm based on IPSAA The concrete steps of HMM tranng algorthm based on IPSAA are as follows: STEP 1: Defne the ftness functon F(x) and ntalze PSO parameters: ncludng populaton sze S, the largest cycle tmes Itermax1, learnng factors c1 and c2, nerta weght W, and random ntalzaton for partcle s poston and velocty wthn the allowable range; // random ntalzaton for HMM parameters; STEP 2: Calculate the functon value of each partcle accordng to Eq.(11). STEP 3: Compare the ftness value of each partcle respectvely wth Indvdual extremum Pbest and global extremum Gbest, and f better, then respectvely substtute, otherwse, reman unchanged; STEP 4: Adaptvely update velocty and poston of partcles accordng to Eq.(1),(2), STEP 5: Restrct and normalze partcles poston;// Restrct and normalze HMM parameters. STEP 6: If the termnaton condton s satsfed (error s good enough or the algorthm reaches PSO s largest cycle DOI /IJSSST.a ISSN: x onlne, prnt

6 tmes Iter max1 ), termnate PSO s optmzaton process, and obtan the best hstory poston of each partcle, otherwse, return to STEP 2; STEP 7: Intalze the ant colony s maxmum cycle tmes termax2 and ants search radus δ, ntalze the poston of ant colony accordng to the optmal hstory poston of each partcle, and ntalze pheromone based on Eq.(9), let j 0,ter2=1, and fnd the best ftness value and the correspondng poston STEP 8: Each ant selects area accordng to the regon probablty decson rules n Eq.(6),(12) and locally searches the huntng spot wthn the radus of δ, f a good soluton s searched locally, then t s replaced, and then the pheromones of huntng pont s adaptvely ncreased accordng to the equaton (7), (8), (9) and (13). STEP 9: Update the optmal ftness value and the correspondng poston. STEP 10: Expand the search radus of ants, ter2 ++, f ter2 <termax2, then go to STEP 8; STEP 11: Output the optmal soluton after ACA s fne search; STEP 12: Take the above IPSAA optmzaton soluton as the nput parameters of Baum-Welch, and locally revse B-W algorthm to obtan the fnal HMM parameter results. Fgure. 2. Flowchart of IPSAA Parameter Tranng The flowchart of IPSAA algorthm s llustrated n Fg.(2). IV. PSO Intalzaton Update partcles poston and velocty Update the ndvdual extreme, local extreme and ther postons Constrant normalzaton recursve teraton Generate the Hstorcal optmal soluton of each partcle Intalze ACA parameters, generate the ntal pheromone dstrbuton based on the optmzed soluton Ants select regon accordng to probablstc decson rule Fnd the huntng pont accordng to Feasble soluton s generaton l Update Pheromone Output HMM parameters after IPSAA optmzaton Baum-Welch algorthm s local revson Output the fnal HMM parameters recursve teraton WEB INFORMATION EXTRACTION BASED ON IPSAA-HMM A. Extracton process based on IPSAA-HMM Ths paper bulds an mproved IPSAA-HMM model. By Usng ctatons n Web research papers as treatment objects, the model extracts state domans n reference such as <Author> <Book> <Ttle> <Journal>. The extracton process of the mproved model s as follows: (1)Informaton preprocessng. Ths artcle frst uses the peweb tool from webste for Web ctaton record extracton; and then utlzes delmter such as punctuaton and text features for nformaton chunkng pretreatment. Among them, the determnstc text features are characterstc word Journal correspondng to journal state doman <Journal>, characterstc words Conference, Proceedngs and Symposum correspondng to conference proceedngs state doman <Conference> and characterstc words Press, Publshers correspondng to press state doman < Press > and so on. (2)Model tranng. After ntalng HMM parameters randomly, the IPSAA algorthm s adopted to optmze HMM parameters and then BW algorthm s used to modfy HMM parameters locally, whch bulds an mproved HMM. (3)Informaton extracton. Vterb algorthm s employed to obtan the optmal state sequence of test sample. The specfc extracton process s shown n Fg.(3). Fgure 3. Web extracton process. B. Expermental results and analyss 2800 unlabeled research paper ctatons are used as expermental samples, a part of whch are 800 ctaton data sets ( from the Unted States Carnege Mellon Unversty (CMU), the other part of whch are 2000 lterature records from 398 research papers extracted randomly from onlne journal database. We select 1900 ctaton records as tranng sets, totalng 45,102 words, the other 900 ctaton records as open test sets, totalng 16,104 words. In HMM optmzaton tranng process, the hybrd tranng parameters are as follows: the populaton sze S=30,Iter max=200, the ntal nerta weght ω nt=0.5, the control parameter of self-adaptve nerta weght n=1.25, the learnng factors c 1 =c 2 =2, the maxmum volatle constrant coeffcent 0 1, the postve coeffcent of adjustng, φ=2.5, the mnmum pheromone mn , max 900, Q=1, m 30, 05., BW teraton threshold ε=1e-5, the HMM state number N=11, the ntal value of λ s selected randomly. We use PSO-HMM, ACA-HMM and IPSAA- BW algorthm n ths paper to tran HMM and analyss the convergence of three algorthms, ther samplng error formula s defned n (14). DOI /IJSSST.a ISSN: x onlne, prnt

7 N M 1 2 ( po ( )) ( po ˆ ( ) po ( )) N M 1 j1 (14) TABLE.2. AVERAGE FΒ=1 COMPARISON OF THREE ALGORITHMS Where ( po ˆ ( ) s the probablty of tranng model, po ( ) s the probablty of sample generaton model. The comparson results are llustrated n Fg.(4). standard devatons PSO-BW ACA-BW IPSAA-BW generatons Fgure 4. Standard devtaton comparson of three algorthms As shown n Fg.(4), the standard error of PSO-HMM and ACA-HMM begn to converge close to 0.18 and 0.11, respectvely, whle that of IPSAA-BW begns to converge only close to 0.04, ts standard error reduces respectvely by about 14% and 7% compared wth the prevous two algorthms. It proves that the mproved algorthm has stronger search ablty, convergence speed and very low error, can more accurately tran HMM model, so as to mprove system qualty. At the same tme, Fg.5 shows the mproved algorthm has better stablty. The nformaton extracton results of three optmzaton algorthms are shown n Table.1. TABLE.1. State EXTRACTION PRECISION AND RECALL COMPARISON OF THREE ALGORITHMS PSO-HMM ACA-HMM IPSAA-HMM Precson Recall Precson Recall Precson Recall Author Ttle Book Journal Conference Press Cty Volume No Year Pages Average PSO-HMM ACA-HMM IPSAA-HMM Fβ= TABLE.3. TIME PERFORMANCE COMPARISON OF THREE ALGORITHMS PSO-HMM ACA-HMM IPSAA-HMM t/s As shown n Table.1, 2, the extracton precson and recall of IPSAA-HMM are all much hgher than the prevous two algorthms. Measurng from the average comprehensve ndex F β=1 value, IPSAA-HMM ncreases respectvely by 8.5% and 6.1% than the prevous two. The precson and recall of states <Journal>, <press> and <conference> ncrease sgnfcantly, manly because the combnaton of determnstc feature nformaton wth hybrd optmzaton model enhances extracton performance. The nterference of state felds such as <book> versus <ttle>,<journal> versus <conference> s strongest, but due to text features and the mproved HMM, the precson rates of states <ttle>, <Journal> and the recall rates of states <book>,<conference> are greatly enhanced. Whle the nterference of <Author> s smaller, ts accuracy rate reached Table.3 shows that the mproved hybrd algorthm has the better tme performance especally than ACA, whch s manly because the new algorthm can avod the blndness n early phrase of ACA and can quckly converge, whch also reflects the effcency of the hybrd algorthm. Syntheszng the above data, t can be seen that the valdty of IPSAA-HMM algorthm. V. CONCLUSIONS In vew of the defects of the tradtonal HMM hybrd method for Web nformaton extracton, ths paper proposes a self-adaptve hybrd ntellgence tranng algorthm based on IPSAA-BW for ctaton extracton. The IPSAA-BW tranng algorthm adjusts adaptvely parameters of PSO and ACA, and takes advantage of the PSO s strong global search capablty to generate the ntal nformaton dstrbuton (Rough search), and then uses the ACA s postve feedback mechansms to obtan the exact solutons (fne search) thus greatly mprovng the performance of HMM parameter optmzaton. And then the algorthm uses BW to revse parameters locally, whch consders the nfluence of the nformaton contaned n tranng sequence and socal nformaton on HMM global optmzaton. So the new hybrd algorthm enhances the probablty of model global optmzaton, effectvely overcomes premature, quckly converges wth extremely low error and has stronger optmzaton ablty. Expermental results show that compared wth the tradtonal PSO-HMM and ACA-HMM optmzaton method, IPSAA-HMM reflects the strong advantage n optmal performance and extracton accuracy, DOI /IJSSST.a ISSN: x onlne, prnt

8 whch proves the effectveness of the mproved algorthm. Future researches can focus on: use new hybrd ntellgent algorthms to development HMM optmzaton method wth better performance and lower complexty, and then apply t to the actual ntellgent nformaton processng system. ACKNOWLEDGEMENTS We would lke to thank to the revewers for ther helpful comments. Ths work was fnancally supported by the Natural Scence Foundaton of Chna (# ), the Hgher School Scence and Technology Development Project n Shanx Provnce of Chna(# ), and the key dscplne constructon project of Xnzhou Teachers Unversty (# XK201403). REFERENCES [1] X. Chen, T. Fang., H. Huo, D.R. L, Measurng the Effectveness of Varous Features for Thematc Informaton Extracton From Very Hgh Resoluton Remote Sensng Imagery, IEEE Transactons On Geoscence And Remote Sensng, vol.53,pp ,2015 ]2] M. Marcnczuk, M. Paseck, Study on named entty recognton for polsh based on hdden Markov models,proceedngs of Text, Speech and Dalogue-13th Internatonal Conference (TSD 2010),pp ,2010. [3] K.R. L, Z.K. Kong, G.X. Chen, and J.W. Zhu, Research on mproved HMM-based text categorzaton, Mcroelectroncs & Computer, vol.29,no.11,pp , [4] M. Zhang, P.Yn, and Z.H. Deng, SVM+BHMM : A Hybrd Statstcal Model for Meta Data Extracton, Journal of Software, vol.19,no.2,pp ,2008. [5] Y. P. Ln, Y.Z. Lu, S. X. Zhou, Z. P. Chen, and L. J. Ca,, Usng hdden Msrkov model for text nformaton extracton based on maxmum entropy, Acta Electronca snca,.vol.33,no.2,pp ,2005 [6] J. Y. Xao, L.M. Zou, and C. Q. L, Hybrd genetc algorthm and hdden Markov model for web nformaton extracton, Computer Engneerng and Applcatons, vol.44,no.18,pp ,2008 [7] L.M. Zou., X. J. Gong, F. Xao, and S. P. Ma, Web nformaton extracton based on smulated annealng algorthm and hdden Markov model, Journal of Unversty of South Chna,,vol.25,no.1,pp.70-74,2011 [8] C. Wang, D. Q. Duan, and X. D. Wang, An mproved PSO and HMM algorthm for web nformaton extracton, Journal of Henan Normal Unversty(Natural Scence),vol.38,no.5,pp:65-68,2010. [9] Y.L. Chen, B.L. Zhong, Facal expresson recognton based on HMM and PSO, Computer Engneerng,vol.34,no.13,pp ,2008 [10] S. J. Yang, S. W. Wang, J. Tao, and X. Lu, Mult-objectve optmzaton method based on hybrd swarm ntellgence algorthm, Computer smulaton, vol.29,no.6,pp ,2012. [11] Sh, Y., Eberhart, R. C., Fuzzy self-adaptve partcle swarm optmzaton, In: Proceedngs of the IEEE Congress on Evolutonary Computaton. Pscataway, NJ: IEEE Servce Center, pp ,2001. [12] X. Zhou, Y.H. Lu, J.D. Zhang, T.M. Lu, and D. Zhang, An ant colony based algorthm for overlappng communty detecton n complex networks, Physca A: Statstcal Mechancs and ts Applcatons, vol.427, no.1,pp ,2015 [13] Mah, M., Baykan, OK., and Kodaz, H., A new hybrd method based on Partcle Swarm Optmzaton, Ant Colony Optmzaton and 3-Opt algorthms for Travelng Salesman Problem, Appled Soft Computng, vol.30,pp: , DOI /IJSSST.a ISSN: x onlne, prnt

BioTechnology. An Indian Journal FULL PAPER. Trade Science Inc.

BioTechnology. An Indian Journal FULL PAPER. Trade Science Inc. [Type text] [Type text] [Type text] ISSN : 0974-74 Volume 0 Issue BoTechnology 04 An Indan Journal FULL PAPER BTAIJ 0() 04 [684-689] Revew on Chna s sports ndustry fnancng market based on market -orented

More information

Determining the Optimal Bandwidth Based on Multi-criterion Fusion

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

More information

Learning the Kernel Parameters in Kernel Minimum Distance Classifier

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

More information

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

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

More information

Clustering Algorithm Combining CPSO with K-Means Chunqin Gu 1, a, Qian Tao 2, b

Clustering Algorithm Combining CPSO with K-Means Chunqin Gu 1, a, Qian Tao 2, b Internatonal Conference on Advances n Mechancal Engneerng and Industral Informatcs (AMEII 05) Clusterng Algorthm Combnng CPSO wth K-Means Chunqn Gu, a, Qan Tao, b Department of Informaton Scence, Zhongka

More information

Parallelism for Nested Loops with Non-uniform and Flow Dependences

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

More information

CHAPTER 2 PROPOSED IMPROVED PARTICLE SWARM OPTIMIZATION

CHAPTER 2 PROPOSED IMPROVED PARTICLE SWARM OPTIMIZATION 24 CHAPTER 2 PROPOSED IMPROVED PARTICLE SWARM OPTIMIZATION The present chapter proposes an IPSO approach for multprocessor task schedulng problem wth two classfcatons, namely, statc ndependent tasks and

More information

Network Intrusion Detection Based on PSO-SVM

Network Intrusion Detection Based on PSO-SVM TELKOMNIKA Indonesan Journal of Electrcal Engneerng Vol.1, No., February 014, pp. 150 ~ 1508 DOI: http://dx.do.org/10.11591/telkomnka.v1.386 150 Network Intruson Detecton Based on PSO-SVM Changsheng Xang*

More information

Problem Definitions and Evaluation Criteria for Computational Expensive Optimization

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

More information

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

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

More information

The Research of Support Vector Machine in Agricultural Data Classification

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

More information

Tsinghua University at TAC 2009: Summarizing Multi-documents by Information Distance

Tsinghua 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 information

Using Particle Swarm Optimization for Enhancing the Hierarchical Cell Relay Routing Protocol

Using Particle Swarm Optimization for Enhancing the Hierarchical Cell Relay Routing Protocol 2012 Thrd Internatonal Conference on Networkng and Computng Usng Partcle Swarm Optmzaton for Enhancng the Herarchcal Cell Relay Routng Protocol Hung-Y Ch Department of Electrcal Engneerng Natonal Sun Yat-Sen

More information

An Optimal Algorithm for Prufer Codes *

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

More information

Kent State University CS 4/ Design and Analysis of Algorithms. Dept. of Math & Computer Science LECT-16. Dynamic Programming

Kent State University CS 4/ Design and Analysis of Algorithms. Dept. of Math & Computer Science LECT-16. Dynamic Programming CS 4/560 Desgn and Analyss of Algorthms Kent State Unversty Dept. of Math & Computer Scence LECT-6 Dynamc Programmng 2 Dynamc Programmng Dynamc Programmng, lke the dvde-and-conquer method, solves problems

More information

Research of Dynamic Access to Cloud Database Based on Improved Pheromone Algorithm

Research of Dynamic Access to Cloud Database Based on Improved Pheromone Algorithm , pp.197-202 http://dx.do.org/10.14257/dta.2016.9.5.20 Research of Dynamc Access to Cloud Database Based on Improved Pheromone Algorthm Yongqang L 1 and Jn Pan 2 1 (Software Technology Vocatonal College,

More information

Maximum Variance Combined with Adaptive Genetic Algorithm for Infrared Image Segmentation

Maximum Variance Combined with Adaptive Genetic Algorithm for Infrared Image Segmentation Internatonal Conference on Logstcs Engneerng, Management and Computer Scence (LEMCS 5) Maxmum Varance Combned wth Adaptve Genetc Algorthm for Infrared Image Segmentaton Huxuan Fu College of Automaton Harbn

More information

An Image Fusion Approach Based on Segmentation Region

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

More information

NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS

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

More information

Recommended Items Rating Prediction based on RBF Neural Network Optimized by PSO Algorithm

Recommended Items Rating Prediction based on RBF Neural Network Optimized by PSO Algorithm Recommended Items Ratng Predcton based on RBF Neural Network Optmzed by PSO Algorthm Chengfang Tan, Cayn Wang, Yuln L and Xx Q Abstract In order to mtgate the data sparsty and cold-start problems of recommendaton

More information

Support Vector Machines

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

More information

Quality Improvement Algorithm for Tetrahedral Mesh Based on Optimal Delaunay Triangulation

Quality Improvement Algorithm for Tetrahedral Mesh Based on Optimal Delaunay Triangulation Intellgent Informaton Management, 013, 5, 191-195 Publshed Onlne November 013 (http://www.scrp.org/journal/m) http://dx.do.org/10.36/m.013.5601 Qualty Improvement Algorthm for Tetrahedral Mesh Based on

More information

FINDING IMPORTANT NODES IN SOCIAL NETWORKS BASED ON MODIFIED PAGERANK

FINDING IMPORTANT NODES IN SOCIAL NETWORKS BASED ON MODIFIED PAGERANK FINDING IMPORTANT NODES IN SOCIAL NETWORKS BASED ON MODIFIED PAGERANK L-qng Qu, Yong-quan Lang 2, Jng-Chen 3, 2 College of Informaton Scence and Technology, Shandong Unversty of Scence and Technology,

More information

THE PATH PLANNING ALGORITHM AND SIMULATION FOR MOBILE ROBOT

THE PATH PLANNING ALGORITHM AND SIMULATION FOR MOBILE ROBOT Journal of Theoretcal and Appled Informaton Technology 30 th Aprl 013. Vol. 50 No.3 005-013 JATIT & LLS. All rghts reserved. ISSN: 199-8645 www.jatt.org E-ISSN: 1817-3195 THE PATH PLANNING ALGORITHM AND

More information

Complexity Analysis of Problem-Dimension Using PSO

Complexity Analysis of Problem-Dimension Using PSO Proceedngs of the 7th WSEAS Internatonal Conference on Evolutonary Computng, Cavtat, Croata, June -4, 6 (pp45-5) Complexty Analyss of Problem-Dmenson Usng PSO BUTHAINAH S. AL-KAZEMI AND SAMI J. HABIB,

More information

Cluster Analysis of Electrical Behavior

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

More information

A Notable Swarm Approach to Evolve Neural Network for Classification in Data Mining

A Notable Swarm Approach to Evolve Neural Network for Classification in Data Mining A Notable Swarm Approach to Evolve Neural Network for Classfcaton n Data Mnng Satchdananda Dehur 1, Bjan Bhar Mshra 2 and Sung-Bae Cho 1 1 Soft Computng Laboratory, Department of Computer Scence, Yonse

More information

Chinese Word Segmentation based on the Improved Particle Swarm Optimization Neural Networks

Chinese Word Segmentation based on the Improved Particle Swarm Optimization Neural Networks Chnese Word Segmentaton based on the Improved Partcle Swarm Optmzaton Neural Networks Ja He Computatonal Intellgence Laboratory School of Computer Scence and Engneerng, UESTC Chengdu, Chna Department of

More information

OPTIMIZATION OF FUZZY RULE BASES USING CONTINUOUS ANT COLONY SYSTEM

OPTIMIZATION 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 information

Design of Structure Optimization with APDL

Design of Structure Optimization with APDL Desgn of Structure Optmzaton wth APDL Yanyun School of Cvl Engneerng and Archtecture, East Chna Jaotong Unversty Nanchang 330013 Chna Abstract In ths paper, the desgn process of structure optmzaton wth

More information

SLAM Summer School 2006 Practical 2: SLAM using Monocular Vision

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

More information

Problem Set 3 Solutions

Problem Set 3 Solutions Introducton to Algorthms October 4, 2002 Massachusetts Insttute of Technology 6046J/18410J Professors Erk Demane and Shaf Goldwasser Handout 14 Problem Set 3 Solutons (Exercses were not to be turned n,

More information

A Time-driven Data Placement Strategy for a Scientific Workflow Combining Edge Computing and Cloud Computing

A Time-driven Data Placement Strategy for a Scientific Workflow Combining Edge Computing and Cloud Computing > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 1 A Tme-drven Data Placement Strategy for a Scentfc Workflow Combnng Edge Computng and Cloud Computng Bng Ln, Fangnng

More information

A New Approach For the Ranking of Fuzzy Sets With Different Heights

A 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 information

An Improved Particle Swarm Optimization for Feature Selection

An Improved Particle Swarm Optimization for Feature Selection Journal of Bonc Engneerng 8 (20)?????? An Improved Partcle Swarm Optmzaton for Feature Selecton Yuannng Lu,2, Gang Wang,2, Hulng Chen,2, Hao Dong,2, Xaodong Zhu,2, Sujng Wang,2 Abstract. College of Computer

More information

Meta-heuristics for Multidimensional Knapsack Problems

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

More information

Parameters Optimization of SVM Based on Improved FOA and Its Application in Fault Diagnosis

Parameters Optimization of SVM Based on Improved FOA and Its Application in Fault Diagnosis Parameters Optmzaton of SVM Based on Improved FOA and Its Applcaton n Fault Dagnoss Qantu Zhang1*, Lqng Fang1, Sca Su, Yan Lv1 1 Frst Department, Mechancal Engneerng College, Shjazhuang, Hebe Provnce,

More information

A Clustering Algorithm Solution to the Collaborative Filtering

A Clustering Algorithm Solution to the Collaborative Filtering Internatonal Journal of Scence Vol.4 No.8 017 ISSN: 1813-4890 A Clusterng Algorthm Soluton to the Collaboratve Flterng Yongl Yang 1, a, Fe Xue, b, Yongquan Ca 1, c Zhenhu Nng 1, d,* Hafeng Lu 3, e 1 Faculty

More information

EECS 730 Introduction to Bioinformatics Sequence Alignment. Luke Huan Electrical Engineering and Computer Science

EECS 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 information

A Fast Content-Based Multimedia Retrieval Technique Using Compressed Data

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

More information

Application of Improved Fish Swarm Algorithm in Cloud Computing Resource Scheduling

Application of Improved Fish Swarm Algorithm in Cloud Computing Resource Scheduling , pp.40-45 http://dx.do.org/10.14257/astl.2017.143.08 Applcaton of Improved Fsh Swarm Algorthm n Cloud Computng Resource Schedulng Yu Lu, Fangtao Lu School of Informaton Engneerng, Chongqng Vocatonal Insttute

More information

Performance Evaluation of Information Retrieval Systems

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

More information

A Novel Adaptive Descriptor Algorithm for Ternary Pattern Textures

A Novel Adaptive Descriptor Algorithm for Ternary Pattern Textures A Novel Adaptve Descrptor Algorthm for Ternary Pattern Textures Fahuan Hu 1,2, Guopng Lu 1 *, Zengwen Dong 1 1.School of Mechancal & Electrcal Engneerng, Nanchang Unversty, Nanchang, 330031, Chna; 2. School

More information

Positive Semi-definite Programming Localization in Wireless Sensor Networks

Positive 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 information

CHAPTER 4 OPTIMIZATION TECHNIQUES

CHAPTER 4 OPTIMIZATION TECHNIQUES 48 CHAPTER 4 OPTIMIZATION TECHNIQUES 4.1 INTRODUCTION Unfortunately no sngle optmzaton algorthm exsts that can be appled effcently to all types of problems. The method chosen for any partcular case wll

More information

K-means Optimization Clustering Algorithm Based on Hybrid PSO/GA Optimization and CS validity index

K-means Optimization Clustering Algorithm Based on Hybrid PSO/GA Optimization and CS validity index Orgnal Artcle Prnt ISSN: 3-6379 Onlne ISSN: 3-595X DOI: 0.7354/jss/07/33 K-means Optmzaton Clusterng Algorthm Based on Hybrd PSO/GA Optmzaton and CS valdty ndex K Jahanbn *, F Rahmanan, H Rezae 3, Y Farhang

More information

Neural Network Based Algorithm for Multi-Constrained Shortest Path Problem

Neural Network Based Algorithm for Multi-Constrained Shortest Path Problem Neural Network Based Algorthm for Mult-Constraned Shortest Path Problem Jyang Dong 1,2, Junyng Zhang 2, and Zhong Chen 1 1 Department of Physcs, Fujan Engneerng Research Center for Sold-State Lghtng, Xamen

More information

Natural Computing. Lecture 13: Particle swarm optimisation INFR /11/2010

Natural Computing. Lecture 13: Particle swarm optimisation INFR /11/2010 Natural Computng Lecture 13: Partcle swarm optmsaton Mchael Herrmann mherrman@nf.ed.ac.uk phone: 0131 6 517177 Informatcs Forum 1.42 INFR09038 5/11/2010 Swarm ntellgence Collectve ntellgence: A super-organsm

More information

BIN XIA et al: AN IMPROVED K-MEANS ALGORITHM BASED ON CLOUD PLATFORM FOR DATA MINING

BIN XIA et al: AN IMPROVED K-MEANS ALGORITHM BASED ON CLOUD PLATFORM FOR DATA MINING An Improved K-means Algorthm based on Cloud Platform for Data Mnng Bn Xa *, Yan Lu 2. School of nformaton and management scence, Henan Agrcultural Unversty, Zhengzhou, Henan 450002, P.R. Chna 2. College

More information

Virtual Memory. Background. No. 10. Virtual Memory: concept. Logical Memory Space (review) Demand Paging(1) Virtual Memory

Virtual Memory. Background. No. 10. Virtual Memory: concept. Logical Memory Space (review) Demand Paging(1) Virtual Memory Background EECS. Operatng System Fundamentals No. Vrtual Memory Prof. Hu Jang Department of Electrcal Engneerng and Computer Scence, York Unversty Memory-management methods normally requres the entre process

More information

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

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

More information

Biostatistics 615/815

Biostatistics 615/815 The E-M Algorthm Bostatstcs 615/815 Lecture 17 Last Lecture: The Smplex Method General method for optmzaton Makes few assumptons about functon Crawls towards mnmum Some recommendatons Multple startng ponts

More information

An efficient iterative source routing algorithm

An efficient iterative source routing algorithm An effcent teratve source routng algorthm Gang Cheng Ye Tan Nrwan Ansar Advanced Networng Lab Department of Electrcal Computer Engneerng New Jersey Insttute of Technology Newar NJ 7 {gc yt Ansar}@ntedu

More information

A Binarization Algorithm specialized on Document Images and Photos

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

More information

An Entropy-Based Approach to Integrated Information Needs Assessment

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

More information

A Fast Visual Tracking Algorithm Based on Circle Pixels Matching

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

More information

Optimizing Document Scoring for Query Retrieval

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

More information

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

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

More information

A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS

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

More information

Analysis of Particle Swarm Optimization and Genetic Algorithm based on Task Scheduling in Cloud Computing Environment

Analysis of Particle Swarm Optimization and Genetic Algorithm based on Task Scheduling in Cloud Computing Environment Analyss of Partcle Swarm Optmzaton and Genetc Algorthm based on Tas Schedulng n Cloud Computng Envronment Frederc Nzanywayngoma School of Computer and Communcaton Engneerng Unversty of Scence and Technology

More information

Research of Neural Network Classifier Based on FCM and PSO for Breast Cancer Classification

Research of Neural Network Classifier Based on FCM and PSO for Breast Cancer Classification Research of Neural Network Classfer Based on FCM and PSO for Breast Cancer Classfcaton Le Zhang 1, Ln Wang 1, Xujewen Wang 2, Keke Lu 2, and Ajth Abraham 3 1 Shandong Provncal Key Laboratory of Network

More information

The Comparison of Calibration Method of Binocular Stereo Vision System Ke Zhang a *, Zhao Gao b

The Comparison of Calibration Method of Binocular Stereo Vision System Ke Zhang a *, Zhao Gao b 3rd Internatonal Conference on Materal, Mechancal and Manufacturng Engneerng (IC3ME 2015) The Comparson of Calbraton Method of Bnocular Stereo Vson System Ke Zhang a *, Zhao Gao b College of Engneerng,

More information

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

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

More information

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

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

More information

Smoothing Spline ANOVA for variable screening

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

More information

Support Vector Machines

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

More information

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

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

More information

An Iterative Solution Approach to Process Plant Layout using Mixed Integer Optimisation

An 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 information

S1 Note. Basis functions.

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

More information

Classifier Swarms for Human Detection in Infrared Imagery

Classifier Swarms for Human Detection in Infrared Imagery Classfer Swarms for Human Detecton n Infrared Imagery Yur Owechko, Swarup Medasan, and Narayan Srnvasa HRL Laboratores, LLC 3011 Malbu Canyon Road, Malbu, CA 90265 {owechko, smedasan, nsrnvasa}@hrl.com

More information

Active Contours/Snakes

Active Contours/Snakes Actve Contours/Snakes Erkut Erdem Acknowledgement: The sldes are adapted from the sldes prepared by K. Grauman of Unversty of Texas at Austn Fttng: Edges vs. boundares Edges useful sgnal to ndcate occludng

More information

Simulation Based Analysis of FAST TCP using OMNET++

Simulation 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 information

Applying Continuous Action Reinforcement Learning Automata(CARLA) to Global Training of Hidden Markov Models

Applying Continuous Action Reinforcement Learning Automata(CARLA) to Global Training of Hidden Markov Models 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

More information

Classifier Selection Based on Data Complexity Measures *

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

More information

An Improved Image Segmentation Algorithm Based on the Otsu Method

An Improved Image Segmentation Algorithm Based on the Otsu Method 3th ACIS Internatonal Conference on Software Engneerng, Artfcal Intellgence, Networkng arallel/dstrbuted Computng An Improved Image Segmentaton Algorthm Based on the Otsu Method Mengxng Huang, enjao Yu,

More information

Straight Line Detection Based on Particle Swarm Optimization

Straight Line Detection Based on Particle Swarm Optimization Sensors & ransducers 013 b IFSA http://www.sensorsportal.com Straght Lne Detecton Based on Partcle Swarm Optmzaton Shengzhou XU, Jun IE College of computer scence, South-Central Unverst for Natonaltes,

More information

Available online at Available online at Advanced in Control Engineering and Information Science

Available 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 information

Image Feature Selection Based on Ant Colony Optimization

Image Feature Selection Based on Ant Colony Optimization Image Feature Selecton Based on Ant Colony Optmzaton Lng Chen,2, Bolun Chen, Yxn Chen 3, Department of Computer Scence, Yangzhou Unversty,Yangzhou, Chna 2 State Key Lab of Novel Software Tech, Nanng Unversty,

More information

ON SOME ENTERTAINING APPLICATIONS OF THE CONCEPT OF SET IN COMPUTER SCIENCE COURSE

ON SOME ENTERTAINING APPLICATIONS OF THE CONCEPT OF SET IN COMPUTER SCIENCE COURSE Yordzhev K., Kostadnova H. Інформаційні технології в освіті ON SOME ENTERTAINING APPLICATIONS OF THE CONCEPT OF SET IN COMPUTER SCIENCE COURSE Yordzhev K., Kostadnova H. Some aspects of programmng educaton

More information

Collaboratively Regularized Nearest Points for Set Based Recognition

Collaboratively Regularized Nearest Points for Set Based Recognition Academc Center for Computng and Meda Studes, Kyoto Unversty Collaboratvely Regularzed Nearest Ponts for Set Based Recognton Yang Wu, Mchhko Mnoh, Masayuk Mukunok Kyoto Unversty 9/1/013 BMVC 013 @ Brstol,

More information

CS 534: Computer Vision Model Fitting

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

More information

Modular PCA Face Recognition Based on Weighted Average

Modular PCA Face Recognition Based on Weighted Average odern Appled Scence odular PCA Face Recognton Based on Weghted Average Chengmao Han (Correspondng author) Department of athematcs, Lny Normal Unversty Lny 76005, Chna E-mal: hanchengmao@163.com Abstract

More information

Research and Application of Fingerprint Recognition Based on MATLAB

Research and Application of Fingerprint Recognition Based on MATLAB Send Orders for Reprnts to reprnts@benthamscence.ae The Open Automaton and Control Systems Journal, 205, 7, 07-07 Open Access Research and Applcaton of Fngerprnt Recognton Based on MATLAB Nng Lu* Department

More information

X- Chart Using ANOM Approach

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

More information

On Some Entertaining Applications of the Concept of Set in Computer Science Course

On 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

Cost-efficient deployment of distributed software services

Cost-efficient deployment of distributed software services 1/30 Cost-effcent deployment of dstrbuted software servces csorba@tem.ntnu.no 2/30 Short ntroducton & contents Cost-effcent deployment of dstrbuted software servces Cost functons Bo-nspred decentralzed

More information

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

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

More information

OPTIMIZATION OF SKELETAL STRUCTURES USING IMPROVED GENETIC ALGORITHM BASED ON PROPOSED SAMPLING SEARCH SPACE IDEA

OPTIMIZATION OF SKELETAL STRUCTURES USING IMPROVED GENETIC ALGORITHM BASED ON PROPOSED SAMPLING SEARCH SPACE IDEA INTERNATIONAL JOURNAL OF OPTIMIZATION IN CIVIL ENGINEERING Int. J. Optm. Cvl Eng., 2018; 8(3): 415-432 OPTIMIZATION OF SKELETAL STRUCTURES USING IMPROVED GENETIC ALGORITHM BASED ON PROPOSED SAMPLING SEARCH

More information

ENERGY EFFICIENCY OPTIMIZATION OF MECHANICAL NUMERICAL CONTROL MACHINING PARAMETERS

ENERGY EFFICIENCY OPTIMIZATION OF MECHANICAL NUMERICAL CONTROL MACHINING PARAMETERS ENERGY EFFICIENCY OPTIMIZATION OF MECHANICAL NUMERICAL CONTROL MACHINING PARAMETERS Zpeng LI*, Ren SHENG Yellow Rver Conservancy Techncal Insttute, School of Mechancal Engneerng, Henan 475000, Chna. Correspondng

More information

User Authentication Based On Behavioral Mouse Dynamics Biometrics

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

More information

Image Emotional Semantic Retrieval Based on ELM

Image Emotional Semantic Retrieval Based on ELM Internatonal Conference on Logstcs Engneerng, Management and Computer Scence (LEMCS 2014) Image Emotonal Semantc Retreval Based on ELM Pele Zhang, Mn Yao, Shenzhang La College of computer scence & Technology

More information

Module Management Tool in Software Development Organizations

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

More information

Edge Detection in Noisy Images Using the Support Vector Machines

Edge 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 information

VISUAL SELECTION OF SURFACE FEATURES DURING THEIR GEOMETRIC SIMULATION WITH THE HELP OF COMPUTER TECHNOLOGIES

VISUAL SELECTION OF SURFACE FEATURES DURING THEIR GEOMETRIC SIMULATION WITH THE HELP OF COMPUTER TECHNOLOGIES UbCC 2011, Volume 6, 5002981-x manuscrpts OPEN ACCES UbCC Journal ISSN 1992-8424 www.ubcc.org VISUAL SELECTION OF SURFACE FEATURES DURING THEIR GEOMETRIC SIMULATION WITH THE HELP OF COMPUTER TECHNOLOGIES

More information

An Influence of the Noise on the Imaging Algorithm in the Electrical Impedance Tomography *

An Influence of the Noise on the Imaging Algorithm in the Electrical Impedance Tomography * Open Journal of Bophyscs, 3, 3, 7- http://dx.do.org/.436/ojbphy.3.347 Publshed Onlne October 3 (http://www.scrp.org/journal/ojbphy) An Influence of the Nose on the Imagng Algorthm n the Electrcal Impedance

More information

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

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

More information

Multi-objective Virtual Machine Placement for Load Balancing

Multi-objective Virtual Machine Placement for Load Balancing Mult-obectve Vrtual Machne Placement for Load Balancng Feng FANG and Bn-Bn Qu,a School of Computer Scence & Technology, Huazhong Unversty Of Scence And Technology, Wuhan, Chna Abstract. The vrtual machne

More information

BAYESIAN MULTI-SOURCE DOMAIN ADAPTATION

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

More information

Research on Kruskal Crossover Genetic Algorithm for Multi- Objective Logistics Distribution Path Optimization

Research on Kruskal Crossover Genetic Algorithm for Multi- Objective Logistics Distribution Path Optimization , pp.367-378 http://dx.do.org/.14257/jmue.215..8.36 Research on Kruskal Crossover Genetc Algorthm for Mult- Objectve Logstcs Dstrbuton Path Optmzaton Yan Zhang 1,2, Xng-y Wu 1 and Oh-kyoung Kwon 2, a,

More information

MULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION

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

More information