Utilizing Content to Enhance a Usage-Based Method for Web Recommendation based on Q-Learning

Size: px
Start display at page:

Download "Utilizing Content to Enhance a Usage-Based Method for Web Recommendation based on Q-Learning"

Transcription

1 Proceedngs of the Twenty-Frst Internatonal FLAIS Conference (2008) Utlzng Content to Enhance a Usage-Based Method for Web ecommendaton based on Q-Learnng Nma Taghpour Department of Computer Engneerng Amrkabr Unversty of Technology, Tehran, Iran n-taghpour@aut.ac.r Abstract The problem of nformaton overload on the Internet has receved a great deal of attenton n the recent years. ecommender Systems have been ntroduced as one soluton to ths problem. These systems am at drectng the user toward the tems that best meet her needs and nterests. ecent studes have ndcated the effectveness of ncorporatng doman knowledge n mprovng the qualty of recommendatons. In ths paper we explot ths approach to enhance a renforcement learnng framework, prmarly devsed for web recommendatons based on web usage data. A hybrd,.e. content- and usage-based, web recommendaton method s proposed by ncorporatng web content nformaton nto a model of user behavor learned form usage data. Content nformaton s utlzed to fnd smlartes between usage scenaros,.e. users' seekng ther nformaton needs, and new recommendaton strateges are proposed that are based on ths enhanced model of user behavor. We evaluate our method under dfferent settngs and show how ths method can overcome the shortcomngs of the usage-based approach and mprove the overall qualty of recommendatons. Introducton The volume of nformaton avalable on the nternet s ncreasng rapdly wth the explosve growth of the World Wde Web and the advent of e-commerce. Whle n one hand, users are provded wth more nformaton and servce optons, on the other hand t has become more dffcult for them to fnd the rght or nterestng nformaton, the problem commonly known as nformaton overload. ecommender systems have been ntroduced as a soluton to ths problem (esnck and aran 1997). They can be generally defned as systems that gude users toward nterestng or useful objects n a large space of possble optons (Burke 2002). One popular applcaton area for recommender systems s web content recommendatons. Web recommendaton s consdered a user modelng or web personalzaton task (Ernak et al. 2004). One research area that has recently contrbuted greatly to ths problem s web mnng. Most of Copyrght 2008, Assocaton for the Advancement of Artfcal Intellgence ( All rghts reserved. the systems developed n ths feld are based on web usage mnng (Srvastava et al. 2000) whch s the process of applyng data mnng technques to the dscovery of usage patterns form web data. These systems are manly concerned wth dscoverng patterns from web usage logs and makng recommendatons based on the extracted navgaton patterns (Fu et al. 2000; Mobasher et al. 2000a). Unlke tradtonal recommender systems, whch manly base ther decsons on user ratngs on dfferent tems or other explct feedbacks provded by the user, these technques dscover user preferences from ther mplct feedbacks, namely the web pages they have vsted. More recently, hybrd methods that take advantage of doman knowledge,.e. content, usage and even structural nformaton of the webstes, have been ntroduced (Bose et al. 2006; Ernak et al. 2003; L & Zaane 2004; Mobasher et al. 2000b; Nakagawa & Mobasher 2003) and shown superor results n the web page recommendaton problem. In (Nakagawa & Mobasher, 2003) the degree of connectvty based on the lnk structure of the webste s used to evaluate effectveness of dfferent usage based technques for web stes wth dfferent structures. A new method for generatng navgaton models s presented n (L & Zaane 2004) whch explots the usage, content and structure data of the webste and addresses parallel nformaton needs of the user. Ernak et al. (2004, 2003) use the content of web pages to augment usage profles wth semantcs usng a doman-ontology. Most recently, concept herarches were ncorporated n a novel recommendaton method based on web usage mnng and optmal sequence algnment to fnd conceptual smlartes between user sessons (Bose et al. 2006). In ths paper we explot ths dea to enhance a renforcement learnng soluton, devsed for web recommendatons based on web usage data (Taghpour et al. 2007). Although the mentoned technque has shown promsng results n comparson to common technques lke collaboratve flterng and assocaton rules, an analyss of the system's performance, showed that ths method suffers from the problems commonly faced by other usage-based technques. We tackle these problems by proposng a hybrd,.e. content- and usage-based, web recommendaton method by explotng web content nformaton n the model of user behavor learned form 101

2 usage data. We devse content models for user navgaton sequence and utlze the content nformaton to fnd the regulartes and smlartes between usage scenaros,.e. users' seekng ther nformaton needs by browsng the web. New recommendaton strateges are proposed based on ths enhanced model of user behavor. Content-wse smlar usage scenaros are exploted as cases of user nformaton need, based on ther sequental nteractons wth web content, whch also ndcate what tems should be recommended to satsfy the nformaton need. Our hybrd model for the web page recommendaton problem emphaszes the flexblty of the renforcement learnng framework for ths problem and how t can be utlzed to ncorporate other sources of nformaton. We evaluate our method under dfferent settngs and show how ths method can mprove the overall qualty of web recommendatons. The organzaton of the paper s as follows: Frst, we overvew the usage-based method whch s the bass of our method. Next, we gve detaled descrptons of our hybrd methods. Afterwards we evaluate the methods and fnally comes our concluson along wth recommendatons for future work. Background In ths secton we overvew the method proposed n (Taghpour et al. 2007) that forms the bass of our new soluton. The proposed method explots enforcement Learnng (L) to make recommendatons from web usage data. We also pont out the man weaknesses of the method whch we am to overcome. Web ecommendatons Based on enforcement Learnng enforcement learnng (Sutton & Barto 1998) s prmarly known n machne learnng research as a framework n whch agents learn to choose the optmal acton n each stuaton or state they are n. The goal of the agent s to learn whch actons to perform n each state to receve the greatest accumulatve reward, n ts path to the goal state. To model the problem as renforcement learnng, they use the analogy of a game n whch the system s constantly tryng to predct the next pages of the user sesson, knowng her prevous requests and the hstory of other users browsng sessons. Usng the notons of N-Grams, each state S at tme t conssts of two sequences, ndcatng the sequence of last w vsted and w' recommended pages respectvely: s =< pt =< s w+ t w + 1 1, pt w+ 2,..., pt > (1),,..., > t w + 2 t Where P and ndcate the th vsted and recommended page n the state (Fgure 1). eward for each acton would be a functon of s' and s' where S ' s the next state. A state S ' s rewarded when the last page vsted belongs to the recommended pages lst. To completely defne the reward common metrcs normally used n web page recommender systems are taken nto account. One aspect to consder s when the vsted page was actually predcted by the system, n order to reward recommendatons that shorten the browsng sessons. Another factor commonly consdered n theses systems (Mobasher et al. 2000a; Fu et al. 2000) s the tme the user spends on a page. The common assumpton s that the more tme the user spends on a page the more nterested he probably s n that page. The rewardng can be summarzed as: Algorthm 1: Usage Based eward Functon 1: Assume δ ( s, a ) = s 2: P = s, w s = Pt + 1 s 3: If P Ø 4: For each page r n P 5: r(s,a,s') =r(s',p t+1 ) +=reward(dst( s, r),tme(p t+1 )) 6: End For 7: End If Where Dst(s, r) s the dstance of page r from the end of the recommended pages lst, and Tme(P t+1 ) s the tme user has spent on the last page of the state. As the sequence of prevously recommended pages s restrcted to a constant number w', the effect of each acton s lmted to w' next states and the system was mostly successful n recommendng pages vsted around w' steps ahead. Ths tends to lmt system s predcton ablty as large numbers of w' make the state space enormous. To overcome ths problem a modfcaton s devsed n reward functon. The basc dea s that when an acton/recommendaton s approprate n state S, ndcatng the recommended page s lkely to occur n the followng states, t should also be consdered approprate n state S -1, the actons n S -1 that frequently lead to S. Lmtatons of the Usage-Based Approach In our evaluaton of the system, we notced that although we were faced wth a rather large number of states, there were cases where the state resulted from the sequence of pages vsted by the user had actually never occurred n the tranng phase. Although not the case here, ths problem can be also due to the nfamous "new tem" problem commonly faced n collaboratve flterng (Burke 2002; Mobasher et al. 2000b) when new pages are added to the webste. In stuatons lke these the system was unable to make any decsons regardng what pages to recommend. S 1 <-,-,a> c S 2 <-,a,b> d S 3 <a,b,c> e S 4 <b,c,d> s S 5 <c,d,e> f. <-,-,-> r c <-,-,c> r d <-,c,d> r e <c,d,e> r s <d,e,s> r f Fgure 1: States and actons n the recommendaton model 102

3 Moreover, the overall coverage of the system on the webste,.e. percentage of the pages that were recommended at least once, was rather low (55.06%). Another ssue worth consderng s the fact that the mere presence of a state n the state space cannot guarantee a hgh qualty recommendaton,.e. a hgh Q-value cannot guarantee a hgh qualty recommendaton by tself. Smply put, when a pattern has few occurrences n the tranng data t cannot be a strong bass for decson makng, a problem addressed n other methods by ntroducng metrcs lke support threshold n assocaton rules (Mobasher et al. 2000b). Smlarly n our case a hgh Q-value, lke a hgh confdence for an assocaton rule, cannot be trusted unless t has strong supportng evdence n the data. Generally, there are cases when hstorcal usage data provdes no evdence, or evdence that's not strong enough, to make a ratonal decson about user's behavor. Ths s a problem common n recommender systems that have usage data as ther only source of nformaton. Note that n the descrbed settng, pages stored n the sequence of each state S are treated as tems for whch the only nformaton avalable s ther d. The system reles solely on usage data and thus s unable to make any generalzaton. One common soluton to ths problem s to ncorporate some semantc knowledge about the tems beng recommended, nto the system. In the next secton we descrbe our approach for adoptng ths dea Explotng Content Implcatons of the Usage- Based Navgatonal Model Motvaton In the usage-based model of the problem, each state models the nformaton need of the user as the (sub)sequence of web pages vsted by the user and the nterest he has shown n each page, e.g. by the tme spent vstng the page. Snce no addtonal nformaton besde the page d s used n the sequence, ths method fals to recognze any new sequence of page vsts and consequently fals to make predctons. The same s true regardng the acton/recommendatons, as the system would only be able to recommend the exact pages t has seen before. In ths secton we wll elaborate our approach to address these ssues based on usng the content of the pages n each state. We devse content models for the states and explot these models to fnd the states representng smlar user nformaton needs. Havng ths knowledge, we wll be able to make use of the learned user behavor n satsfyng smlar nformaton needs. For ths purpose we devse new recommendaton strateges that make use of clusters of smlar states and the approprate recommendatons, for those states to derve a new aggregate score for predctons, n a k-nn fashon. The basc dea s that whenever the user browsng results n a state S x whch has been vsted no or few tmes before, the system would fnd a set Sm k (S x ) contanng the k most smlar states to S x and compute new recommendaton scores for each canddate web page r by utlzng the scores,.e. Q-values, of recommendng r n each state S Sm k (S x ). For example the aggregaton can be done by a weghted sum of the Q-alues, such as: Score ( S x, r) = S Sm k ( S x ) W ( S ) Q( S, r) S Sm k ( S x ) W ( S ) In the followng sectons we wll present dfferent strateges for the recommendaton phase. Smlarty between the States Fndng smlartes between states s not a trval task. Frst, we need to model the content nformaton encapsulated n the state. Any state contans a fxed number (w) of sequental page vsts by a user. The content nformaton of ths sequence can be modeled n varous ways e.g. w bag of words, a vector aggregatng all page contents n a state, etc. We present two approaches to model the content nformaton of the states n attempt to capture a model of user nformaton need based on her traversal on the webste: Content Sequence Model (CSM). In ths approach we explot varous sources of nformaton about each web page p j n the sequence, combne ths nformaton nto an aggregated vector PC(p j ) n the vector space model and fnally derve a content model SC(S ) for each state S as an ordered set of these w vectors. The content vector of each web page s computed by combnng the content of the web page, the terms n the UL of the page and the anchor text of the hyperlnks pontng to the web page InLnk(p j ). Ths model s adopted from (Ernak et al. 2003): PC( p ) = α. Content( p ) + β. InLnk( p ) + γ. UL( p )) (3) j j j j = ( t 1, t2, t3,..., tn), where α + β + γ = 1 SC S ) =< PC ( p ), PC ( p ),..., PC ( p ) > (4) ( 1 2 w Where N s the total number of terms as commonly consdered n the vector space model and the weghts t j are computed usng the tf.idf metrc. Havng defned the content model the next step s to devse the smlarty functon CSM_Sm(S,S j ). We compute ths smlarty as the weghted cosne-based smlarty (CosSm) of correspondng pages n the states. More weght s assgned to pages that appear later n the user sesson based on the assumpton that users are browsng towards ther nformaton need. The smlarty s computed as: w k = 1 CosSm( PC( p ), PC( p )) pweght( k) w k = 1 k jk pweght( k) Informaton Scent Model (ISM). Informaton Scent (Ch et. al. 2001) s a model of user nformaton need based on (5) (2) 103

4 the nformaton foragng theory. In ths model user's nformaton need s modeled usng the text of the hyperlnks followed by the user n her browsng sesson on the webste and applyng a spreadng actvaton algorthm. We refer the nterested reader to (Ch et. al. 2001) for detals of ths approach. We explot ths algorthm to acheve a model of user nformaton need based on the sequence of vsted pages n each state. The result of ths procedure s also a vector n the vector space model and the smlarty between states wll agan be computed based on the cosne of the correspondng vectors. Organzaton of Smlar States Another ssue worth mentonng s that regardless of the choce of the content model, the process of fndng smlar states has to be tmeeffcent as ths process s performed durng the onlne recommendaton generaton phase. In ths regard, the states are clustered based on ther content model n the offlne tranng phase and the search space for fndng k smlar states wll be reduced to the correspondng cluster of the gven state S x. We ncorporated our smlarty functons and modfed the Dc-tree clusterng algorthm (Wong and Fu 2000) for formng state clusters n our experments. Ths algorthm s an ncremental herarchcal clusterng algorthm specfcally devsed for fndng clusters of web pages. The ncremental feature of the algorthm specfcally desred as t enables us to accommodate new states nto the approprate cluster when needed. ecommendaton Generaton The performance of the system reles heavly on the recommendaton strategy. We propose two approaches for web page recommendaton by explotng the content nformaton of the navgaton model, each wth dfferent motvatons. ecommendaton based on an Aggregate of Smlar States (ASS). Ths strategy s analogous to the approach mentoned as the motvaton. Here, we explot the fact that the usage-based model generates accurate recommendatons when t s provded wth suffcent usage data (Taghpour et al. 2007). So, whenever the user sesson results n a state S u whch has been frequently vsted by prevous users, recommendatons are made solely based on the values of Q(S u,a) for dfferent actons. On the other hand, the usage-based method loses ts accuracy when faced wth less frequently vsted sequence of pages s completely useless when faced wth new sequences. So, whenever S u s a new or rarely vsted state, an aggregate predcton of smlar states wll be exploted for recommendaton. The pseudo code n Algorthm 2 summarzes the procedure. If S u s a new state the correspondng content model SC(S u ) of the state s computed. Ths model s then compared to the mean content vector of the state clusters and the nearest cluster s found based on Equaton (5). In ths smlarty computaton, pweght(k) s also dependant to the tme the user has spent on each pages p uk, assumng pages the user has spent more tme on as stronger ndcators of her nterest. In the last step recommendaton scores for canddate recommendatons/actons wll be computed usng the nearest states S Sm k (S u ) n the selected cluster and by applyng Equaton (2). Note that n ths step we tend to assgn more weght (W(S )) to the states whch have occurred more frequently n the usage data. The whole procedure s the same where S u resdes n our state space but t has been vsted less than a gven threshold. The only dfference s that n ths case we know the cluster the state belongs to n advance and there's no need to compute the correspondng content model and fnd nearest clusters. Algorthm 2: ASS 1: Assume Current State s S u 2: If (S u s not a new state) 3: If Usage_Support(S u ) > mn_sup 4: eturn BestActons(S u ) 5: Ext; 6: Else 7: C FndCluster(S u ) 8: End If 9: Else 10: Buld SC(S u ) 11: C ClosestCluster(SC(S u )) 12: End If 13: Sm k (S u ) FndSmlar(C, SC(S u ),k) 14: For each Acton r 15: Compute Score(S u,r) usng Equaton (2) 16: Store Score(S u,r) n a set ecs 17: End For 18: eturn BestActons(ecs) Usng ths strategy, we wll be able to fnd portons of sessons whch are both semantcally smlar to our sesson, n sense of the user nformaton need, and have a rather stronger usage support. Then we recommend a combnaton of actons approprate for those states. We expect to be able to cover a hgher porton of user sessons and also gan hgher recommendaton coverage on the web pages. Although ths method manages the new sequences (even those contanng new pages added to the webste) t stll fals to recommend those newly added pages (at least before a deal of expermentaton). Inferrng a Content-based Model for ecommendatons from Smlar States (CMSS). In ths strategy, agan we explot the Q-alues of actons n smlar state to a gven state S u, but ths tme nstead of usng the actons as they are,.e. predcton of page ds, we derve an aggregate content model predctng the content of pages that mght satsfy user nformaton needs. The procedure s bascally smlar to Algorthm 2. Agan a set ecs s found contanng the best actons from states n Sm k (S u ). Afterwards, the content model of the top m pages, those wth hgher Score(S u,r) n ecs, wll be used to derve an aggregate content model of pages that mght satsfy the user's nformaton need. Then ths content model (ecq), as a weghted vector of terms, wll be used as a query on pages of the webste. The ranked web pages retreved for the query wll then be used as the recommendatons to be 104

5 presented to the user. ecq s computed as normalzed weghted sum of the content vectors of pages n ecs: e cq = r Top m (e cs ) PC ( r ) Score ( S r Top m (e cs ) Score ( S u, r ) u, r ) Ths strategy s also a hybrd of usage and content-based approaches, and t dffers from content-based recommendatons where content smlar to the tems accessed so far by the user s retreved for recommendatons. Here, the prevous usage patterns are exploted and the content query used for recommendatons s based on a predcton of the content that would follow the sequence of web pages accessed by the user. An mportant feature of ths strategy s the possblty of recommendng new pages added to the web ste, as the content queres would be evaluated aganst all the web pages, ncludng the new pages or even pages wth dynamc content. Expermental Evaluaton Expermental Settng We evaluated system performance n dfferent settngs descrbed above. As our evaluaton data set we used the web logs of the unversty webste. Ths dataset contans sessons and about 680 pages. 60% of the data set was used as the tranng set and the remanng was used to test the system. For our evaluaton we presented each user sesson to the system, and recorded the recommendatons t made after seeng each page the user had vsted smlar to the orgnal approach presented n (Taghpour et al. 2007). We used the metrcs proposed n (Bose et al. 2006) for evaluaton. These metrcs are: Ht ato (H): Percentage of Hts n recommendaton lsts, Predctve Ablty (PA): Percentage of pages recommended at least once, Clck educton (C): average percentage of pages skpped because of recommendatons and ecommendaton Qualty (Q): average rank of a correct recommendaton n the lst. Senstvty to Parameters Dfferent choce of parameters would result n dfferent types of system performance. mn_sup (Algorthm2) would determne when recommendatons for the exstng states are made based on the usage-based approach, and when smlar states would be exploted. We evaluated the usagebased method and analyzed the relatonshp between the frequences of each state's sequence n tranng sessons and the average accuracy of recommendatons made for the state. These results showed sgnfcance reducton of accuracy for states wth support values lower than 0.31%, the rato that we chose as the mn-sup threshold. Another parameter s k the sze of the S x neghborhood Sm k (S x ). We evaluated the system wth dfferent k values as shown n Fgure 2. We have a local optmum for k n both strateges (k=15 for ASS and k=20 fro CMSS). Increasng k at lower (6) value ranges mproves the performance by brngng n the knowledge stored n smlar states. After reachng a local optmum the performance stays rather flat as we obtan an almost fxed set of best actons, wth hghest Score(S u,r). Increasng the neghborhood beyond 30 seems to brng nose nto the model whch hurts the performance n spte of the dampenng effect of low W(S ) values for the less smlar neghbors. Average Ht ato ( H) Avg. Predctve Ablty (PA) ASS CMSS Neghborhood Sze (k ) ASS CMSS UBL Neghborhood Sze (k ) Fgure 2: Effect of Parameter k on Ht ato (Top) and Predctve Ablty (Bottom) of the Method Comparson to Other Methods We compared our proposed hybrd methods wth the prevous usage-based (UB-L) and a content-based approach that uses the nfo-scent model to recommend pages (CIS). Note that UB-L had shown superor results than common usage-based methods (Taghpour et al. 2007). Combnaton of the chosen strategy (ASS/CMSS) and content model (CSM/ISM) result n 4 dfferent methods. The results presented here are based on havng a maxmum of 7 recommendatons n each stage. We also expermented wth 5 and 10 as the thresholds whch resulted n the same relatve performance of the methods. As show n Table 1, the ASS-CSM out performs the rest of the methods (ncludng the accurate UB-L method) n sense of H, whle also achevng hgh levels of coverage on the webste. We can see how all our hybrd methods manage to dramatcally outperform UB-L n sense of PA and farly compete wth the, not so accurate, content-based CIS method. Both of the methods based on the CMSS strategy acheve hghest levels of PA (and C), due to the dversty of the content queres. The much hgher H values of these methods compared to the pure contentbased CIS approach could be consdered as evdence n support of the mportance of actual usage patterns n accurate nference of user nformaton needs and behavor. 105

6 Table 1: Comparson of dfferent recommendaton methods Sesson wndow sze w=3, mn_sup=0.31%, k=15 (ASS), k=20 (CMSS) Method Metrc H PA C Q PA-N UB-L CIS ASS-CSM ASS-ISM CMSS-CSM CMSS-ISM Fnally, we can see that the more elaborate CSM content model results n relatvely better performance than the ISM approach, but the margnal dfference proves the accuracy of the ISM approach, whch s favorable due to less computaton load, especally for the onlne recommendaton phase. We also evaluated the ablty of the methods to recommend new pages, omtted from tranng sessons n another smulaton, shown by the evaluaton metrc PA-N. It can be seen that the CMSS strateges manage to acheve a hgh coverage on these pages whle other approaches fal, as expected. Concluson and Future Work We presented methods to enhance, and overcome the restrctons, of a usage-based web recommender system based on L. We dentfed and tackled the weaknesses that a usage-based method nherently suffers from and ncorporated content nformaton regardng the usage patterns to mprove the system. Our evaluaton results emphasze the flexblty of the L paradgm to combne dfferent sources of nformaton n order to derve predctons of future content needs and mprove the qualty of recommendatons. Ths work can be extended n varous dmensons. One possblty s a recommendaton method that uses a combnaton of recommendatons generated by the two strateges, n order to beneft from each,.e. better accuracy of the ASS, dversty of CMSS and ts ablty to accommodate new pages. Another opton s usng a content-based, or a hybrd, reward functon n the L model whch consders the smlarty of pages n to the vsted page. Integraton of other sources of doman knowledge e.g. webste topology or a doman-ontology nto the model can also be another future work for ths paper. Fnally, we propose ntegraton of some long lastng characterstcs of the user, e.g. her long-term nterests or a model of her browsng strategy, nto the state model. eferences Bose, A., Beemanapall, K., Srvastava, J., Sahar, S Incorporatng concept herarches nto usage mnng based recommendatons. Proc. 8th WEBKDD workshop. Burke, Hybrd recommender systems: Survey and experments. User Modelng and User-Adapted Interacton. 12: Ch, E. H., Proll, P., Ptkow, J Usng nformaton scent to model user Informaton Needs and Actons on the Web. Proc.s of Human Factors n Computng Systems. Ernak, M., azrganns, M., arlams, I SEWeP: Usng Ste Semantcs and a Taxonomy to Enhance the Web Personalzaton Process. Proc. of the 9th ACM SIGKDD Conference. Ernak, M., Lampos, C., Paulaks, S., azrganns, M Web Personalzaton Integratng Content Semantcs and Navgatonal Patterns. Proc. of the sxth ACM workshop on Web Informaton and Data Management. Fu, X., Budzk, J., Hammond, K. J Mnng navgaton hstory for recommendaton. Proc.of 4th Internatonal Conference on Intellgent User Interfaces. L, J., Zaane, O Combnng usage, content and structure data to mprove web ste recommendaton. 5th Internatonal Conference on Electronc Commerce and Web. EC-WEB. Mobasher, B., Cooley,., Srvastava, J Automatc personalzaton based on web usage mnng. Communcatons of the ACM. 43 (8), pp Mobasher, B., Da, H., Luo, T., Sun, Y., Zhu, J Integratng web usage and content mnng for more effectve personalzaton. In Proc. of Internatonal Conference on Electronc Commerce and Web. EC-WEB. Nakagawa M., Mobasher, B A Hybrd Web Personalzaton Model Based on Ste Connectvty. Proc. of 5th WEBKDD workshop. esnck, P., aran, H ecommender Systems. Communcatons of the ACM, 40 (3): Srvastava, J., Cooley,., Deshpande, M., Tan, P.N Web Usage Mnng: Dscovery and Applcatons of Usage Patterns from Web Data. SIGKDD Exploratons, 1: Sutton,.S. Barto, A.G enforcement Learnng: An Introducton, MIT Press, Cambrdge. Taghpour, N., Kardan, A., Shry Ghdary, S Usage- Based Web ecommendatons: A enforcement Learnng Approach. Proc. of the ACM 2007 ecommender Systems Conference. Wong, W. and Fu, A Incremental document clusterng for web page classfcaton. Proc. Internatonal conference on Informaton Socety n the 21st century: Emergng Technologes and New Challenges. 106

Effective Page Recommendation Algorithms Based on. Distributed Learning Automata and Weighted Association. Rules

Effective Page Recommendation Algorithms Based on. Distributed Learning Automata and Weighted Association. Rules Effectve Page Recommendaton Algorthms Based on Dstrbuted Learnng Automata and Weghted Assocaton Rules R. Forsat 1*, M. R. Meybod 2 1 Department of Computer Engneerng, Islamc Azad Unversty, Karaj Branch,

More information

Query Clustering Using a Hybrid Query Similarity Measure

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

More information

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

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

More information

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

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

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

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

More information

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

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

More information

A Webpage Similarity Measure for Web Sessions Clustering Using Sequence Alignment

A Webpage Similarity Measure for Web Sessions Clustering Using Sequence Alignment A Webpage Smlarty Measure for Web Sessons Clusterng Usng Sequence Algnment Mozhgan Azmpour-Kv School of Engneerng and Scence Sharf Unversty of Technology, Internatonal Campus Ksh Island, Iran mogan_az@ksh.sharf.edu

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

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

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

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

More information

LinkSelector: A Web Mining Approach to. Hyperlink Selection for Web Portals

LinkSelector: A Web Mining Approach to. Hyperlink Selection for Web Portals nkselector: A Web Mnng Approach to Hyperlnk Selecton for Web Portals Xao Fang and Olva R. u Sheng Department of Management Informaton Systems Unversty of Arzona, AZ 8572 {xfang,sheng}@bpa.arzona.edu Submtted

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

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

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

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

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

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

Enhancement of Infrequent Purchased Product Recommendation Using Data Mining Techniques

Enhancement of Infrequent Purchased Product Recommendation Using Data Mining Techniques Enhancement of Infrequent Purchased Product Recommendaton Usng Data Mnng Technques Noraswalza Abdullah, Yue Xu, Shlomo Geva, and Mark Loo Dscplne of Computer Scence Faculty of Scence and Technology Queensland

More information

UB at GeoCLEF Department of Geography Abstract

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

More information

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

The Effect of Similarity Measures on The Quality of Query Clusters

The Effect of Similarity Measures on The Quality of Query Clusters The effect of smlarty measures on the qualty of query clusters. Fu. L., Goh, D.H., Foo, S., & Na, J.C. (2004). Journal of Informaton Scence, 30(5) 396-407 The Effect of Smlarty Measures on The Qualty of

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

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

Wishing you all a Total Quality New Year!

Wishing you all a Total Quality New Year! Total Qualty Management and Sx Sgma Post Graduate Program 214-15 Sesson 4 Vnay Kumar Kalakband Assstant Professor Operatons & Systems Area 1 Wshng you all a Total Qualty New Year! Hope you acheve Sx sgma

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

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

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

More information

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

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

Machine Learning 9. week

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

More information

Hierarchical clustering for gene expression data analysis

Hierarchical clustering for gene expression data analysis Herarchcal clusterng for gene expresson data analyss Gorgo Valentn e-mal: valentn@ds.unm.t Clusterng of Mcroarray Data. Clusterng of gene expresson profles (rows) => dscovery of co-regulated and functonally

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

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

Load Balancing for Hex-Cell Interconnection Network

Load Balancing for Hex-Cell Interconnection Network Int. J. Communcatons, Network and System Scences,,, - Publshed Onlne Aprl n ScRes. http://www.scrp.org/journal/jcns http://dx.do.org/./jcns.. Load Balancng for Hex-Cell Interconnecton Network Saher Manaseer,

More 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

Concurrent Apriori Data Mining Algorithms

Concurrent Apriori Data Mining Algorithms Concurrent Apror Data Mnng Algorthms Vassl Halatchev Department of Electrcal Engneerng and Computer Scence York Unversty, Toronto October 8, 2015 Outlne Why t s mportant Introducton to Assocaton Rule Mnng

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

Feature-Based Matrix Factorization

Feature-Based Matrix Factorization Feature-Based Matrx Factorzaton arxv:1109.2271v3 [cs.ai] 29 Dec 2011 Tanq Chen, Zhao Zheng, Quxa Lu, Wenan Zhang, Yong Yu {tqchen,zhengzhao,luquxa,wnzhang,yyu}@apex.stu.edu.cn Apex Data & Knowledge Management

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

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

Mining Web Logs with PLSA Based Prediction Model to Improve Web Caching Performance

Mining Web Logs with PLSA Based Prediction Model to Improve Web Caching Performance JOURAL OF COMPUTERS, VOL. 8, O. 5, MAY 2013 1351 Mnng Web Logs wth PLSA Based Predcton Model to Improve Web Cachng Performance Chub Huang Department of Automaton, USTC Key laboratory of network communcaton

More information

Alignment Results of SOBOM for OAEI 2010

Alignment Results of SOBOM for OAEI 2010 Algnment Results of SOBOM for OAEI 2010 Pegang Xu, Yadong Wang, Lang Cheng, Tany Zang School of Computer Scence and Technology Harbn Insttute of Technology, Harbn, Chna pegang.xu@gmal.com, ydwang@ht.edu.cn,

More information

Mathematics 256 a course in differential equations for engineering students

Mathematics 256 a course in differential equations for engineering students Mathematcs 56 a course n dfferental equatons for engneerng students Chapter 5. More effcent methods of numercal soluton Euler s method s qute neffcent. Because the error s essentally proportonal to the

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

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

Signed Distance-based Deep Memory Recommender

Signed Distance-based Deep Memory Recommender Sgned Dstance-based Deep Memory Recommender ABSTRACT Personalzed recommendaton algorthms learn a user s preference for an tem, by measurng a dstance/smlarty between them. However, some of exstng recommendaton

More information

Solving two-person zero-sum game by Matlab

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

More information

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

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

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

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

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

More information

User Tweets based Genre Prediction and Movie Recommendation using LSI and SVD

User Tweets based Genre Prediction and Movie Recommendation using LSI and SVD User Tweets based Genre Predcton and Move Recommendaton usng LSI and SVD Saksh Bansal, Chetna Gupta Department of CSE/IT Jaypee Insttute of Informaton Technology,sec-62 Noda, Inda sakshbansal76@gmal.com,

More information

Steps for Computing the Dissimilarity, Entropy, Herfindahl-Hirschman and. Accessibility (Gravity with Competition) Indices

Steps for Computing the Dissimilarity, Entropy, Herfindahl-Hirschman and. Accessibility (Gravity with Competition) Indices Steps for Computng the Dssmlarty, Entropy, Herfndahl-Hrschman and Accessblty (Gravty wth Competton) Indces I. Dssmlarty Index Measurement: The followng formula can be used to measure the evenness between

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

Feature Selection as an Improving Step for Decision Tree Construction

Feature Selection as an Improving Step for Decision Tree Construction 2009 Internatonal Conference on Machne Learnng and Computng IPCSIT vol.3 (2011) (2011) IACSIT Press, Sngapore Feature Selecton as an Improvng Step for Decson Tree Constructon Mahd Esmael 1, Fazekas Gabor

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

A Hybrid Genetic Algorithm for Routing Optimization in IP Networks Utilizing Bandwidth and Delay Metrics

A Hybrid Genetic Algorithm for Routing Optimization in IP Networks Utilizing Bandwidth and Delay Metrics A Hybrd Genetc Algorthm for Routng Optmzaton n IP Networks Utlzng Bandwdth and Delay Metrcs Anton Redl Insttute of Communcaton Networks, Munch Unversty of Technology, Arcsstr. 21, 80290 Munch, Germany

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

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

Machine Learning: Algorithms and Applications

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

More information

An Approach for Recommender System by Combining Collaborative Filtering with User Demographics and Items Genres

An Approach for Recommender System by Combining Collaborative Filtering with User Demographics and Items Genres Volume 18 No.13, October 015 An Approach for Recommender System by Combnng Collaboratve Flterng wth User Demographcs and Items Genres Saurabh Kumar Twar Department of Informaton Technology Samrat Ashok

More information

A Background Subtraction for a Vision-based User Interface *

A Background Subtraction for a Vision-based User Interface * A Background Subtracton for a Vson-based User Interface * Dongpyo Hong and Woontack Woo KJIST U-VR Lab. {dhon wwoo}@kjst.ac.kr Abstract In ths paper, we propose a robust and effcent background subtracton

More information

Recommender System based on Higherorder Logic Data Representation

Recommender System based on Higherorder Logic Data Representation Recommender System based on Hgherorder Logc Data Representaton Lnna L Bngru Yang Zhuo Chen School of Informaton Engneerng, Unversty Of Scence and Technology Bejng Abstract Collaboratve Flterng help users

More information

Personalized Information Retrieval Approach

Personalized Information Retrieval Approach Personalzed Informaton Retreval Approach Myram Hadjoun 1, Mohamed Ramz Haddad 1, Hajer Baazaou 1, Mare-Aude Aufaure 2, and Henda Ben Ghezala 1 1 Laboratory RIADI-GDL, Natonal School of Computer Scences,

More information

Constructing Minimum Connected Dominating Set: Algorithmic approach

Constructing Minimum Connected Dominating Set: Algorithmic approach Constructng Mnmum Connected Domnatng Set: Algorthmc approach G.N. Puroht and Usha Sharma Centre for Mathematcal Scences, Banasthal Unversty, Rajasthan 304022 usha.sharma94@yahoo.com Abstract: Connected

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

Incremental Learning with Support Vector Machines and Fuzzy Set Theory

Incremental Learning with Support Vector Machines and Fuzzy Set Theory The 25th Workshop on Combnatoral Mathematcs and Computaton Theory Incremental Learnng wth Support Vector Machnes and Fuzzy Set Theory Yu-Mng Chuang 1 and Cha-Hwa Ln 2* 1 Department of Computer Scence and

More 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

Document Representation and Clustering with WordNet Based Similarity Rough Set Model

Document Representation and Clustering with WordNet Based Similarity Rough Set Model IJCSI Internatonal Journal of Computer Scence Issues, Vol. 8, Issue 5, No 3, September 20 ISSN (Onlne): 694-084 www.ijcsi.org Document Representaton and Clusterng wth WordNet Based Smlarty Rough Set Model

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

Related-Mode Attacks on CTR Encryption Mode

Related-Mode Attacks on CTR Encryption Mode Internatonal Journal of Network Securty, Vol.4, No.3, PP.282 287, May 2007 282 Related-Mode Attacks on CTR Encrypton Mode Dayn Wang, Dongda Ln, and Wenlng Wu (Correspondng author: Dayn Wang) Key Laboratory

More information

A Deflected Grid-based Algorithm for Clustering Analysis

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

More information

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

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

More information

An Evolvable Clustering Based Algorithm to Learn Distance Function for Supervised Environment

An Evolvable Clustering Based Algorithm to Learn Distance Function for Supervised Environment IJCSI Internatonal Journal of Computer Scence Issues, Vol. 7, Issue 5, September 2010 ISSN (Onlne): 1694-0814 www.ijcsi.org 374 An Evolvable Clusterng Based Algorthm to Learn Dstance Functon for Supervsed

More information

Your Neighbors Affect Your Ratings: On Geographical Neighborhood Influence to Rating Prediction

Your Neighbors Affect Your Ratings: On Geographical Neighborhood Influence to Rating Prediction Your Neghbors Affect Your Ratngs: On Geographcal Neghborhood Influence to Ratng Predcton Longke Hu lhu003@e.ntu.edu.sg Axn Sun axsun@ntu.edu.sg Yong Lu luy0054@e.ntu.edu.sg School of Computer Engneerng,

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

Complex Numbers. Now we also saw that if a and b were both positive then ab = a b. For a second let s forget that restriction and do the following.

Complex Numbers. Now we also saw that if a and b were both positive then ab = a b. For a second let s forget that restriction and do the following. Complex Numbers The last topc n ths secton s not really related to most of what we ve done n ths chapter, although t s somewhat related to the radcals secton as we wll see. We also won t need the materal

More information

Abstract. 1 Introduction

Abstract. 1 Introduction Challenges and Solutons for Synthess of Knowledge Regardng Collaboratve Flterng Algorthms Danel Lowd, Olver Godde, Matthew McLaughln, Shuzhen Nong, Yun Wang, and Jonathan L. Herlocker. School of Electrcal

More information

CORE: A Tool for Collaborative Ontology Reuse and Evaluation

CORE: A Tool for Collaborative Ontology Reuse and Evaluation CRE: A Tool for Collaboratve ntology Reuse and Evaluaton Mram Fernández, Iván Cantador, Pablo Castells Escuela Poltécnca Superor Unversdad Autónoma de Madrd Campus de Cantoblanco, 28049 Madrd, Span {mram.fernandez,

More information

A Lazy Ensemble Learning Method to Classification

A Lazy Ensemble Learning Method to Classification IJCSI Internatonal Journal of Computer Scence Issues, Vol. 7, Issue 5, September 2010 ISSN (Onlne): 1694-0814 344 A Lazy Ensemble Learnng Method to Classfcaton Haleh Homayoun 1, Sattar Hashem 2 and Al

More information

Virtual Machine Migration based on Trust Measurement of Computer Node

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

More information

Combining Multiple Resources, Evidence and Criteria for Genomic Information Retrieval

Combining Multiple Resources, Evidence and Criteria for Genomic Information Retrieval Combnng Multple Resources, Evdence and Crtera for Genomc Informaton Retreval Luo S 1, Je Lu 2 and Jame Callan 2 1 Department of Computer Scence, Purdue Unversty, West Lafayette, IN 47907, USA ls@cs.purdue.edu

More information

SHAPE RECOGNITION METHOD BASED ON THE k-nearest NEIGHBOR RULE

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

More information

Load-Balanced Anycast Routing

Load-Balanced Anycast Routing Load-Balanced Anycast Routng Chng-Yu Ln, Jung-Hua Lo, and Sy-Yen Kuo Department of Electrcal Engneerng atonal Tawan Unversty, Tape, Tawan sykuo@cc.ee.ntu.edu.tw Abstract For fault-tolerance and load-balance

More information

A MODIFIED K-NEAREST NEIGHBOR CLASSIFIER TO DEAL WITH UNBALANCED CLASSES

A MODIFIED K-NEAREST NEIGHBOR CLASSIFIER TO DEAL WITH UNBALANCED CLASSES A MODIFIED K-NEAREST NEIGHBOR CLASSIFIER TO DEAL WITH UNBALANCED CLASSES Aram AlSuer, Ahmed Al-An and Amr Atya 2 Faculty of Engneerng and Informaton Technology, Unversty of Technology, Sydney, Australa

More information

LRD: Latent Relation Discovery for Vector Space Expansion and Information Retrieval

LRD: Latent Relation Discovery for Vector Space Expansion and Information Retrieval LRD: Latent Relaton Dscovery for Vector Space Expanson and Informaton Retreval Techncal Report KMI-06-09 March, 006 Alexandre Gonçalves, Janhan Zhu, Dawe Song, Vctora Uren, Roberto Pacheco In Proc. of

More information

A Web Recommendation Technique Based on Probabilistic Latent Semantic Analysis

A Web Recommendation Technique Based on Probabilistic Latent Semantic Analysis A Web Recommendaton Technque Based on Probablstc Latent Semantc Analyss Guandong Xu 1, Yanchun Zhang 1, and Xaofang Zhou 2 1 School of Computer Scence and Mathematcs, Vctora Unversty, PO Box 14428, VIC

More information

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

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

More information

Corner-Based Image Alignment using Pyramid Structure with Gradient Vector Similarity

Corner-Based Image Alignment using Pyramid Structure with Gradient Vector Similarity Journal of Sgnal and Informaton Processng, 013, 4, 114-119 do:10.436/jsp.013.43b00 Publshed Onlne August 013 (http://www.scrp.org/journal/jsp) Corner-Based Image Algnment usng Pyramd Structure wth Gradent

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

Determining Fuzzy Sets for Quantitative Attributes in Data Mining Problems

Determining Fuzzy Sets for Quantitative Attributes in Data Mining Problems Determnng Fuzzy Sets for Quanttatve Attrbutes n Data Mnng Problems ATTILA GYENESEI Turku Centre for Computer Scence (TUCS) Unversty of Turku, Department of Computer Scence Lemmnkäsenkatu 4A, FIN-5 Turku

More information

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

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

More information

A User Selection Method in Advertising System

A User Selection Method in Advertising System Int. J. Communcatons, etwork and System Scences, 2010, 3, 54-58 do:10.4236/jcns.2010.31007 Publshed Onlne January 2010 (http://www.scrp.org/journal/jcns/). A User Selecton Method n Advertsng System Shy

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

Data Mining: Model Evaluation

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

More information

Deep Classification in Large-scale Text Hierarchies

Deep Classification in Large-scale Text Hierarchies Deep Classfcaton n Large-scale Text Herarches Gu-Rong Xue Dkan Xng Qang Yang 2 Yong Yu Dept. of Computer Scence and Engneerng Shangha Jao-Tong Unversty {grxue, dkxng, yyu}@apex.sjtu.edu.cn 2 Hong Kong

More information

Feature Reduction and Selection

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

More information

Discovering similar user navigation behavior in Web log data

Discovering similar user navigation behavior in Web log data Dscoverng smlar user navgaton behavor n Web log data Tawfq A. Al-asad 1 and Ahmed J. Obad 2 1 Department of Computer Scence, Drector, College of IT, Babylon Unversty, Iraq. 2 Department of Computer Scence,

More information

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

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

More information

2x x l. Module 3: Element Properties Lecture 4: Lagrange and Serendipity Elements

2x x l. Module 3: Element Properties Lecture 4: Lagrange and Serendipity Elements Module 3: Element Propertes Lecture : Lagrange and Serendpty Elements 5 In last lecture note, the nterpolaton functons are derved on the bass of assumed polynomal from Pascal s trangle for the fled varable.

More information