An Iterative Implicit Feedback Approach to Personalized Search

Size: px
Start display at page:

Download "An Iterative Implicit Feedback Approach to Personalized Search"

Transcription

1 An Iteratve Implct Feedback Approach to Personalzed Search Yuanhua Lv 1, Le Sun 2, Junln Zhang 2, Jan-Yun Ne 3, Wan Chen 4, and We Zhang 2 1, 2 Insttute of Software, Chnese Academy of Scences, Beng, , Chna 3 Unversty of Montreal, Canada 1 lvyuanhua@gmal.com 2 {sunle, unln01, zhangwe04}@scas.cn 3 4 ne@ro.umontreal.ca chenwan@nus.edu.sg Abstract General nformaton retreval systems are desgned to serve all users wthout consderng ndvdual needs. In ths paper, we propose a novel approach to personalzed search. It can, n a unfed way, explot and utlze mplct feedback nformaton, such as query logs and mmedately vewed documents. Moreover, our approach can mplement result re-rankng and query expanson smultaneously and collaboratvely. Based on ths approach, we develop a clent-sde personalzed web search agent PAIR (Personalzed Assstant for Informaton Retreval), whch supports both Englsh and Chnese. Our experments on TREC and HTRDP collectons clearly show that the new approach s both effectve and effcent. 1 Introducton Analyss suggests that, whle current nformaton retreval systems, e.g., web search engnes, do a good ob of retrevng results to satsfy the range of ntents people have, they are not so well n dscernng ndvduals search goals (J. Teevan et al., 2005). Search engnes encounter problems such as query ambguty and results ordered by popularty rather than relevance to the user s ndvdual needs. To overcome the above problems, there have been many attempts to mprove retreval accuracy based on personalzed nformaton. Relevance Feedback (G. Salton and C. Buckley, 1990) s the man post-query method for automatcally mprovng a system s accuracy of a user s ndvdual need. The technque reles on explct relevance assessments (.e. ndcatons of whch documents contan relevant nformaton). Relevance feedback has been proved to be qute effectve for mprovng retreval accuracy (G. Salton and C. Buckley, 1990; J. J. Roccho, 1971). However, searchers may be unwllng to provde relevance nformaton through explctly markng relevant documents (M. Beauleu and S. Jones, 1998). Implct Feedback, n whch an IR system unobtrusvely montors search behavor, removes the need for the searcher to explctly ndcate whch documents are relevant (M. Morta and Y. Shnoda, 1994). The technque uses mplct relevance ndcatons, although not beng as accurate as explct feedback, s proved can be an effectve substtute for explct feedback n nteractve nformaton seekng envronments (R. Whte et al., 2002). In ths paper, we utlze the mmedately vewed documents, whch are the clcked results n the same query, as one type of mplct feedback nformaton. Research shows that relatve preferences derved from mmedately vewed documents are reasonably accurate on average (T. Joachms et al., 2005). Another type of mplct feedback nformaton that we explot s users query logs. Anyone who uses search engnes has accumulated lots of clck through data, from whch we can know what queres were, when queres occurred, and whch search results were selected to vew. These query logs provde valuable nformaton to capture users nterests and preferences. Both types of mplct feedback nformaton above can be utlzed to do result re-rankng and query expanson, (J. Teevan et al., 2005; Xuehua Shen. et al., 2005) whch are the two general approaches to personalzed search. (J. Ptkow et al., 2002) However, to the best of our knowledge, how to explot these two types of mplct feedback n a unfed way, whch not only brngs collaboraton between query expanson and result re-rankng but also makes the whole system more concse, has so far not been well studed n the prevous work. In ths paper, we adopt HITS algorthm (J. Klenberg, 1998), and propose a 585 Proceedngs of the 21st Internatonal Conference on Computatonal Lngustcs and 44th Annual Meetng of the ACL, pages , Sydney, July c 2006 Assocaton for Computatonal Lngustcs

2 HITS-lke teratve approach addressng such a problem. Our work dffers from exstng work n several aspects: (1) We propose a HITS-lke teratve approach to personalzed search, based on whch, mplct feedback nformaton, ncludng mmedately vewed documents and query logs, can be utlzed n a unfed way. (2) We mplement result re-rankng and query expanson smultaneously and collaboratvely trggered by every clck. (3) We develop and evaluate a clent-sde personalzed web search agent PAIR, whch supports both Englsh and Chnese. The remanng of ths paper s organzed as follows. Secton 2 descrbes our novel approach for personalzed search. Secton 3 provdes the archtecture of PAIR system and some specfc technques. Secton 4 presents the detals of the experment. Secton 5 dscusses the prevous work related to our approach. Secton 6 draws some conclusons of our work. 2 Iteratve Implct Feedback Approach We propose a HITS-lke teratve approach for personalzed search. HITS (Hyperlnk-Induced Topc Search) algorthm, frst descrbed by (J. Klenberg, 1998), was orgnally used for the detecton of hgh-score hub and authorty web pages. The Authorty pages are the central web pages n the context of partcular query topcs. The strongest authorty pages conscously do not lnk one another 1 they can only be lnked by some relatvely anonymous hub pages. The mutual renforcement prncple of HITS states that a web page s a good authorty page f t s lnked by many good hub pages, and that a web page s a good hub page f t lnks many good authorty pages. A drected graph s constructed, of whch the nodes represent web pages and the drected edges represent hyperlnks. After teratvely computng based on the renforcement prncple, each node gets an authorty score and a hub score. In our approach, we explot the relatonshps between documents and terms n a smlar way to HITS. Unseen search results, those results whch are retreved from search engne yet not been presented to the user, are consdered as authorty pages. Representatve terms are consdered as hub pages. Here the representatve terms are the terms extracted from and best representng the mplct feedback nformaton. Representatve terms confer a relevance score to the unseen 1 For nstance, There s hardly any other company s Web page lnked from search results specfcally, the unseen search results, whch contan more good representatve terms, have a hgher possblty of beng relevant; the representatve terms should be more representatve, f they occur n the unseen search results that are more lkely to be relevant. Thus, also there s mutual renforcement prncple exstng between representatve terms and unseen search results. By the same token, we constructed a drected graph, of whch the nodes ndcate unseen search results and representatve terms, and the drected edges represent the occurrence of the representatve terms n the unseen search results. The followng Table 1 shows how our approach corresponds to HITS algorthm. Approaches The Drected Graph Nodes Edges HITS Authorty Pages Hub Pages Hyperlnks Our Unseen Search Representatve Occurrence Approach Results Terms 2 Table 1. Our approach versus HITS. Because we have already known that the representatve terms are hub pages, and that the unseen search results are authorty pages, wth respect to the former, only hub scores need to be computed; wth respect to the latter, only authorty scores need to be computed. Fnally, after teratvely computng based on the mutual renforcement prncple we can re-rank the unseen search results accordng to ther authorty scores, as well as select the representatve terms wth hghest hub scores to expand the query. Below we present how to construct a drected graph to begn wth. 2.1 Constructng a Drected Graph We can vew the unseen search results and the representatve terms as a drected graph G = (V, E). A sample drected graph s shown n Fgure 1: Fgure 1. A sample drected graph. The nodes V correspond to the unseen search results (the rectangles n Fgure 1) and the repre- 2 The occurrence of the representatve terms n the unseen search results. 586

3 sentatve terms (the crcles n Fgure 1); a drected edge p q E s weghed by the frequency of the occurrence of a representatve term p n an unseen search result q (e.g., the number put on the edge t 1 r 2 ndcates that t 1 occurs twce n r 2 ). We say that each representatve term only has an out-degree whch s the number of the unseen search results t occurs n, as well as that each unseen search result only has an n-degree whch s the count of the representatve terms t contans. Based on ths, we assume that the unseen search results and the representatve terms respectvely correspond to the authorty pages and the hub pages ths assumpton s used throughout the proposed algorthm. 2.2 A HITS-lke Iteratve Algorthm In ths secton, we present how to ntalze the drected graph and how to teratvely compute the authorty scores and the hub scores. And then accordng to these scores, we show how to re-rank the unseen search results and expand the ntal query. Intally, each unseen search result of the query are consdered equally authortatve, that s, = = = 1 Y (1) y y y 1 2 Y Where vector Y ndcates authorty scores of the overall unseen search results, and Y s the sze of such a vector. Meanwhle, each representatve term, wth the term frequency tf n the hstory query logs that have been udged related to the current query, obtans ts hub score accordng to the follow formulaton: x 0 tf X = 1 tf = (2) Where vector X ndcates hub scores of the overall representatve terms, and X s the sze of the vector X. The nodes of the drected graph are ntalzed n ths way. Next, we assocate each edge wth a weght: w( ) = (3) t r tf, Where tf, ndcates the term frequency of the representatve term t occurrng n the unseen search result r ; w(t r ) s the weght of edge that lnk from t to r. For nstance, n Fgure 1, w(t 1 r 2 ) = 2. After ntalzaton, the teratvely computng of hub scores and authorty scores starts. The hub score of each representatve term s re-computed based on three factors: the authorty scores of each unseen search result where ths term occurs; the occurrng frequency of ths term n each unseen search result; the total occurrence of every representatve term n each unseen search result. The formulaton for re-computng hub scores s as follows: ( 1) ( ) ' k + k w t r (4) x = y : t w( ) r t n r n: t n r Where x`(k+1) s the hub score of a representatve term t after (k+1)th teraton; y k s the authorty score of an unseen search result r after kth teraton; : t r ndcates the set of all unseen search results those t occurs n; n: t n r ndcates the set of all representatve terms those r contans. The authorty score of each unseen search result s also re-computed relyng on three factors: the hub scores of each representatve term that ths search result contans; the occurrng frequency of each representatve term n ths search result; the total occurrence of each representatve term n every unseen search results. The formulaton for re-computng authorty scores s as follows: ( 1) ( ) ' k + k w t r (5) y = x : t w( ) r t r m m: t r m Where y`(k+1) s the authorty score of an unseen k search result r after (k+1)th teraton; x s the hub score of a representatve term t after kth teraton; : t r ndcates the set of all representatve terms those r contans; m: t r m ndcates the set of all unseen search results those t occurs n. After re-computaton, the hub scores and the authorty scores are normalzed to 1. The formulaton for normalzaton s as follows: y y ' and Y X y ' x = = x' x' k k= 1 k= 1 k (6) The teraton, ncludng re-computaton and normalzaton, s repeated untl the changes of the hub scores and the authorty scores are smaller than some predefned threshold θ (e.g ). Specfcally, after each repetton, the changes n authorty scores and hub scores are computed usng the followng formulaton: Y ( 1) ( k k x ) 2 ( ( k 1) k ) c y y x x = + = 1 = 1 (7) The teraton stops f c<θ. Moreover, the teraton wll also stop f repetton has reached a 587

4 predefned tmes k (e.g. 30). The procedure of the teraton s shown n Fgure 2. As soon as the teraton stops, the top n unseen search results wth hghest authorty scores are selected and recommended to the user; the top m representatve terms wth hghest hub scores are selected to expand the orgnal query. Here n s a predefned number (n PAIR system we set n=3, n s gven a small number because usng mplct feedback nformaton s sometmes rsky.) m s determned accordng to the poston of the bggest gap, that s, f t t +1 s bgger than the gap of any other two neghborng ones of the top half representatve terms, then m s gven a value. Furthermore, some of these representatve terms (e.g. top 50% hgh score terms) wll be agan used n the next tme of mplementng the teratve algorthm together wth some newly ncomng terms extracted from the ust now clck. Iterate (T, R, k, θ) T: a collecton of m terms R: a collecton of n search results k: a natural number θ: a predefned threshold Apply (1) to ntalze Y. Apply (2) to ntalze X. Apply (3) to ntalze W. For = 1, 2, k Apply (4) to (X -1, Y -1 ) and obtan X`. Apply (5) to (X -1, Y -1 ) and obtan Y`. Apply (6) to Normalze X` and Y`, and respectvely obtan X and Y. Apply (7) and obtan c. If c<θ, then break. End Return (X, Y). Fgure 2. The HITS-lke teratve algorthm. 3 Implementaton 3.1 System Desgn In ths secton, we present our expermental system PAIR, whch s an IE Browser Helper Obect (BHO) based on the popular Web search engne Google. PAIR has three man modules: Result Retreval module, User Interactons module, and Iteratve Algorthm module. The archtecture s shown n Fgure 3. The Result Retreval module runs n backgrounds and retreves results from search engne. When the query has been expanded, ths module wll use the new keywords to contnue retrevng. The User Interactons module can handle three types of basc user actons: (1) submttng a query; (2) clckng to vew a search result; (3) clckng the Next Page lnk. For each of these actons, the system responds wth: (a) explotng and extractng representatve terms from mplct feedback nformaton; (b) fetchng the unseen search results va Results Retreval module; (c) sendng the representatve terms and the unseen search results to Iteratve Algorthm module. Fgure 3. The archtecture of PAIR. The Iteratve Algorthm module mplements the HITS-lke algorthm descrbed n secton 2. When ths module receves data from User Interactons module, t responds wth: (a) teratvely computng the hub scores and authorty scores; (b) re-rankng the unseen search results and expandng the orgnal query. Some specfc technques for capturng and explotng mplct feedback nformaton are descrbed n the followng sectons. 3.2 Extract Representatve Terms from Query Logs We udge whether a query log s related to the current query based on the smlarty between the query log and the current query text. Here the query log s assocated wth all documents that the user has selected to vew. The form of each query log s as follows <query text><query tme> [clcked documents]* The clcked documents consst of URL, ttle and snppet of every clcked document. The reason why we utlze the query text of the current query but not the search results (ncludng ttle, snppet, etc.) to compute the smlarty, s out of consderaton for effcency. If we had used the search results to determne the smlarty, the computaton could only start once the search engne has returned the search results. In our method, nstead, we can explot query logs whle search engne s dong retrevng. Notce that although our system only utlzes the query logs n the last 24 hours; n practce, we can explot much more because of ts low computaton cost wth respect to the retreval process performed n parallel. 588

5 Google result query = aguar CA - Cars Cars Apple - Mac OS X Apple - Support UK - Cars UK - R s for dspace.dal.ppex.com/ Schrödnger -> Home Schrödnger -> Ste Map query = aguar After the 4 th result beng clcked CA - Cars Cars Apple - Mac OS X Amazon.com: Mac OS X Mac OS X 10.2 arstechnca.com/revews/os Macworld: News: Macworld maccentral.macworld.com/news/ Apple - Support -3 UK - Cars -3 UK - R s for -3 PAIR result query = aguar car query logs UK - Cars UK - R s for CA - Cars -2 Cars -2 Apple - Mac OS X -2 Apple - Support -2 dspace.dal.ppex.com/ Schrödnger -> Home Schrödnger -> Ste Map Table 2. Sample results of re-rankng. The search results n boldface are the ones that our system recommends to the user. -3 and -2 n the rght sde of some results ndcate the how ther ranks descend. We use the standard vector space retreval model (G. Salton and M. J. McGll, 1983) to compute the smlarty. If the smlarty between any query log and the current query exceeds a predefned threshold, the query log wll be consdered to be related to current query. Our system wll attempt to extract some (e.g. 30%) representatve terms from such related query logs accordng to the weghts computed by applyng the followng formulaton: w( t ) = tf df (8) Where tf and df respectvely are the term frequency and nverse document frequency of t n the clcked documents of a related query log. Ths formulaton means that a term s more representatve f t has a hgher frequency as well as a broader dstrbuton n the related query log. 3.3 Extract Representatve Terms from Immedately Vewed Documents The representatve terms extracted from mmedately vewed documents are determned based on three factors: term frequency n the mmedately vewed document, nverse document frequency n the entre seen search results, and a dscrmnant value. The formulaton s as follows: ( ) d r d N w x = ( ) tf df d x x x (9) Where tf dr x s the term frequency of term x n the vewed results set d r ; tf dr x s the nverse document frequency of x n the entre seen results set d N. And the dscrmnant value d(x ) of x s computed usng the weghtng schemes F2 (S. E. Robertson and K. Sparck Jones, 1976) as follows: rr d( x ) ln = (10) ( n r ) ( N R ) Where r s the number of the mmedately vewed documents contanng term x ; n s the number of the seen results contanng term x ; R s the number of the mmedately vewed documents n the query; N s the number of the entre seen results. 3.4 Sample Results Unlke other systems whch do result re-rankng and query expanson respectvely n dfferent ways, our system mplements these two functons smultaneously and collaboratvely Query expanson provdes dversfed search results whch must rely on the use of re-rankng to be moved forward and recommended to the user. Fgure 4. A screen shot for query expanson. After teratvely computng usng our approach, the system selects some search results wth top hghest authorty scores and recommends them to the user. In Table 2, we show that PAIR successfully re-ranks the unseen search results of aguar respectvely usng the mmedately 589

6 vewed documents and the query logs. Smultaneously, some representatve terms are selected to expand the orgnal query. In the query of aguar (wthout query logs), we clck some results about Mac OS, and then we see that a term Mac has been selected to expand the orgnal query, and some results of the new query aguar Mac are recommended to the user under the help of re-rankng, as shown n Fgure 4. 4 Experment 4.1 Expermental Methodology It s a challenge to quanttatvely evaluate the potental performance mprovement of the proposed approach over Google n an unbased way (D. Hawkng et al., 1999; Xuehua Shen et al., 2005). Here, we adopt a smlar quanttatve evaluaton as what Xuehua Shen et al. (2005) do to evaluate our system PAIR and recrut 9 students who have dfferent backgrounds to partcpate n our experment. We use query topcs from TREC 2005 and 2004 Hard Track, TREC 2004 Terabyte track for Englsh nformaton retreval, 3 and use query topcs from HTRDP 2005 Evaluaton for Chnese nformaton retreval. 4 The reason why we utlze multple TREC tasks rather than usng a sngle one s that more queres are more lkely to cover the most nterestng topcs for each partcpant. Intally, each partcpant would freely choose some topcs (typcally 5 TREC topcs and 5 HTRDP topcs). Each query of TREC topcs wll be submtted to three systems: UCAIR 5 (Xuehua Shen et al., 2005), PAIR No QE (PAIR system of whch the query expanson functon s blocked) and PAIR. Each query of HTRDP topcs needs only to be submtted to PAIR No QE and PAIR. We do not evaluate UCAIR usng HTRDP topcs, snce t does not support Chnese. For each query topc, the partcpants use the ttle of the topc as the ntal keyword to begn wth. Also they can form some other keywords by themselves f the ttle alone fals to descrbe some detals of the topc. There s no lmt on how many queres they must submt. Durng each query process, the partcpant may clck to vew some results, ust as n normal web search. Then, at the end of each query, search results from these dfferent systems are randomly and anonymously mxed together so that every par- 3 Text REtreval Conference HTRDP Evaluaton. 5 The latest verson released on November 11, tcpant would not know where a result comes from. The partcpants would udge whch of these results are relevant. At last, we respectvely measure precson at top 5, top 10, top 20 and top 30 documents of these system. 4.2 Results and Analyss Altogether, 45 TREC topcs (62 queres n all) are chosen for Englsh nformaton retreval. 712 documents are udged as relevant from Google search results. The correspondng number of relevant documents from UCAIR, PAIR No QE and PAIR respectvely s: 921, 891 and Fgure 5 shows the average precson of these four systems at top n documents among such 45 TREC topcs. Fgure 5. Average precson for TREC topcs. 45 HTRDP topcs (66 queres n all) are chosen for Chnese nformaton retreval. 809 documents are udged as relevant from Google search results. The correspondng number of relevant documents from PAIR No QE and PAIR respectvely s: 1198 and Fgure 6 shows the average precson of these three systems at top n documents among such 45 HTRDP topcs. Fgure 6. Average precson for HTRDP topcs. PAIR and PAIR No QE versus Google We can see clearly from Fgure 5 and Fgure 6 that the precson of PAIR s mproved a lot comparng wth that of Google n all measure- 590

7 ments. Moreover, the mprovement scale ncreases from precson at top 10 to that of top 30. One explanaton for ths s that the more mplct feedback nformaton generated, the more representatve terms can be obtaned, and thus, the teratve algorthm can perform better, leadng to more precse search results. PAIR No QE also sgnfcantly outperforms Google n these measurements, however, wth query expanson, PAIR can perform even better. Thus, we say that result re-rankng and query expanson both play an mportant role n PAIR. Comparng Fgure 5 wth Fgure 6, one can see that the mprovement of PAIR versus Google n Chnese IR s even larger than that of Englsh IR. One explanaton for ths s that: before mplementng the teratve algorthm, each Chnese search result, ncludng ttle and snppet, s segmented nto words (or phrases). And only the noun, verb and adectve of these words (or phrases) are used n next stages, whereas, we only remove the stop words for Englsh search result. Another explanaton s that there are some Chnese web pages wth the same content. If one of such pages s clcked, then, occasonally some repetton pages are recommended to the user. However, snce PAIR s based on the search results of Google and the nformaton concernng the result pages that PAIR can obtaned s lmted, whch leads to t dffcult to avod the replcatons. PAIR and PAIR No QE versus UCAIR In Fgure 5, we can see that the precson of PAIR No QE s better than that of UCAIR among top 5 and top 10 documents, and s almost the same as that of UCAIR among top 20 and top 30 documents. However, PAIR s much better than UCAIR n all measurements. Ths ndcates that result re-rankng fals to do ts best wthout query expanson, snce the relevant documents n orgnal query are lmted, and only the re-rankng method alone cannot solve the relevant documents sparseness problem. Thus, the query expanson method, whch can provde fresh and relevant documents, can help the re-rankng method to reach an even better performance. Effcency of PAIR The teraton statstc n evaluaton ndcates that the average teraton tmes of our approach s 22 before convergence on condton that we set the threshold θ = The experment shows that the computaton tme of the proposed approach s mperceptble for users (less than 1ms.) 5 Related Work There have been many pror attempts to personalzed search. In ths paper, we focus on the related work dong personalzed search based on mplct feedback nformaton. Some of the exstng studes capture users nformaton need by explotng query logs. For example, M. Speretta and S. Gauch (2005) buld user profles based on actvty at the search ste and study the use of these profles to provde personalzed search results. F. Lu et al. (2002) learn user's favorte categores from hs query hstory. Ther system maps the nput query to a set of nterestng categores based on the user profle and confnes the search doman to these categores. Some studes mprove retreval performance by explotng users browsng hstory (F. Tanudaa and L. Mu, 2002; M. Morta and Y. Shnoda, 1994) or Web communtes (A. Krtkopoulos and M. Sder, 2003; K. Sugyama et al., 2004) Some studes utlze clent sde nteractons, for example, K. Bharat (2000) automatcally dscovers related materal on behalf of the user by servng as an ntermedary between the user and nformaton retreval systems. Hs system observes users nteractng wth everyday applcatons and then antcpates ther nformaton needs usng a model of the task at hand. Some latest studes combne several types of mplct feedback nformaton. J. Teevan et al. (2005) explore rch models of user nterests, whch are bult from both search-related nformaton, such as prevously ssued queres and prevously vsted Web pages, and other nformaton about the user such as documents and emal the user has read and created. Ths nformaton s used to re-rank Web search results wthn a relevance feedback framework. Our work s partly nspred by the study of Xuehua Shen et al. (2005), whch s closely related to ours n that they also explot mmedately vewed documents and short-term hstory queres, mplement query expanson and re-rankng, and develop a clent-sde web search agents that perform eager mplct feedback. However, ther work dffers from ours n three ways: Frst, they use the cosne smlarty to mplement query expanson, and use Roccho formulaton (J. J. Roccho, 1971) to re-rank the search results. Thus, ther query expanson and re-rankng are computed separately and are not so concse and collaboratve. Secondly, ther query expanson s based only on the past queres and s mplemented before the query, whch leads to that 591

8 ther query expanson does not beneft from user s clck through data. Thrdly, they do not compute the relevance of search results and the relatvty of expanded terms n an teratve fashon. Thus, ther approach does not utlze the relaton among search results, among expanded terms, and between search results and expanded terms. 6 Conclusons In ths paper, we studed how to explot mplct feedback nformaton to mprove retreval accuracy. Unlke most prevous work, we propose a novel HITS-lke teratve algorthm that can make use of query logs and mmedately vewed documents n a unfed way, whch not only brngs collaboraton between query expanson and result re-rankng but also makes the whole system more concse. We further propose some specfc technques to capture and explot these two types of mplct feedback nformaton. Usng these technques, we develop a clent-sde web search agent PAIR. Experments n Englsh and Chnese collectons show that our approach s both effectve and effcent. However, there s stll room to mprove the performance of the proposed approach, such as explotng other types of personalzed nformaton, choosng some more effectve strateges to extract representatve terms, studyng the effects of the parameters used n the approach, etc. Acknowledgement We would lke to thank the anonymous revewers for ther helpful feedback and correctons, and to the nne partcpants of our evaluaton experments. Addtonally, ths work s supported by the Natonal Scence Fund of Chna under contact References A. Krtkopoulos and M. Sder, The Compass Flter: Search engne result personalzaton usng Web communtes. In Proceedngs of ITWP, pages D. Hawkng, N. Craswell, P.B. Thstlewate, and D. Harman, Results and challenges n web search evaluaton. Computer Networks, 31(11-16): F. Lu, C. Yu, and W. Meng, Personalzed web search by mappng user queres to categores. In Proceedngs of CIKM, pages F. Tanudaa and L. Mu, Persona: a contextualzed and personalzed web search. HICSS. G. Salton and M. J. McGll, Introducton to Modern Informaton Retreval. McGraw-Hll. G. Salton and C. Buckley, Improvng retreval performance by relevance feedback. Journal of the Amercan Socety for Informaton Scence, 41(4): J. J. Roccho, Relevance feedback n nformaton retreval. In The SMART Retreval System : Experments n Automatc Document Processng, pages Prentce-Hall Inc. J. Klenberg, Authortatve sources n a hyperlnked envronment. ACM, 46(5): J. Ptkow, H. Schutze, T. Cass, R. Cooley, D. Turnbull, A. Edmonds, E. Adar, and T. Breuel, Personalzed search. Communcatons of the ACM, 45(9): J. Teevan, S. T. Dumas, and E. Horvtz, Personalzng search va automated analyss of nterests and actvtes. In Proceedngs of SIGIR, pages K. Bharat, SearchPad: Explct capture of search context to support Web search. Computer Networks, 33(1-6): K. Sugyama, K. Hatano, and M. Yoshkawa, Adaptve Web search based on user profle constructed wthout any effort from user. In Proceedngs of WWW, pages M. Beauleu and S. Jones, Interactve searchng and nterface ssues n the okap best match retreval system. Interactng wth Computers, 10(3): M. Morta and Y. Shnoda, Informaton flterng based on user behavor analyss and best match text retreval. In Proceedngs of SIGIR, pages M. Speretta and S. Gauch, Personalzng search based on user search hstory. Web Intellgence, pages R. Whte, I. Ruthven, and J. M. Jose, The use of mplct evdence for relevance feedback n web retreval. In Proceedngs of ECIR, pages S. E. Robertson and K. Sparck Jones, Relevance weghtng of search terms. Journal of the Amercan Socety for Informaton Scence, 27(3): T. Joachms, L. Granka, B. Pang, H. Hembrooke, and G. Gay, Accurately Interpretng Clckthrough Data as Implct Feedback, In Proceedngs of SIGIR, pages Xuehua Shen, Bn Tan, and Chengxang Zha, Implct User Modelng for Personalzed Search. In Proceedngs of CIKM, pages

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

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

Description of NTU Approach to NTCIR3 Multilingual Information Retrieval

Description of NTU Approach to NTCIR3 Multilingual Information Retrieval Proceedngs of the Thrd NTCIR Workshop Descrpton of NTU Approach to NTCIR3 Multlngual Informaton Retreval Wen-Cheng Ln and Hsn-Hs Chen Department of Computer Scence and Informaton Engneerng Natonal Tawan

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

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

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

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 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

Personalized Concept-Based Clustering of Search Engine Queries

Personalized Concept-Based Clustering of Search Engine Queries IEEE TRANSACTIONS ON JOURNAL NAME, MANUSCRIPT ID 1 Personalzed Concept-Based Clusterng of Search Engne Queres Kenneth Wa-Tng Leung, Wlfred Ng, and Dk Lun Lee Abstract The exponental growth of nformaton

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 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

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

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

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

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

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

Extraction of User Preferences from a Few Positive Documents

Extraction of User Preferences from a Few Positive Documents Extracton of User Preferences from a Few Postve Documents Byeong Man Km, Qng L Dept. of Computer Scences Kumoh Natonal Insttute of Technology Kum, kyungpook, 730-70,South Korea (Bmkm, lqng)@se.kumoh.ac.kr

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

IN recent years, we have been witnessing the explosive

IN recent years, we have been witnessing the explosive IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, VOL. 15, NO. 4, JULY/AUGUST 2003 1 Query Expanson by Mnng User Logs Hang Cu, J-Rong Wen, Jan-Yun Ne, and We-Yng Ma, Member, IEEE Abstract Queres to

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

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

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

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

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

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

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

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

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

Query classification using topic models and support vector machine

Query classification using topic models and support vector machine Query classfcaton usng topc models and support vector machne Deu-Thu Le Unversty of Trento, Italy deuthu.le@ds.untn.t Raffaella Bernard Unversty of Trento, Italy bernard@ds.untn.t Abstract Ths paper descrbes

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

Modeling Hierarchical User Interests Based on HowNet and Concept Mapping

Modeling Hierarchical User Interests Based on HowNet and Concept Mapping Modelng Herarchcal User Interests Based on HowNet and Concept Mappng Yhong L #1, Fang L #2 # Dept. of Computer Scence & Engneerng, Shangha Jao Tong Unversty No.800 Dong Chuan Rd. Shangha 200240, P.R. Chna

More information

Improving Web Image Search using Meta Re-rankers

Improving Web Image Search using Meta Re-rankers VOLUME-1, ISSUE-V (Aug-Sep 2013) IS NOW AVAILABLE AT: www.dcst.com Improvng Web Image Search usng Meta Re-rankers B.Kavtha 1, N. Suata 2 1 Department of Computer Scence and Engneerng, Chtanya Bharath Insttute

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 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

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

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

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

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

More information

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

Efficient Distributed File System (EDFS)

Efficient Distributed File System (EDFS) Effcent Dstrbuted Fle System (EDFS) (Sem-Centralzed) Debessay(Debsh) Fesehaye, Rahul Malk & Klara Naherstedt Unversty of Illnos-Urbana Champagn Contents Problem Statement, Related Work, EDFS Desgn Rate

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

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

Keyword-based Document Clustering

Keyword-based Document Clustering Keyword-based ocument lusterng Seung-Shk Kang School of omputer Scence Kookmn Unversty & AIrc hungnung-dong Songbuk-gu Seoul 36-72 Korea sskang@kookmn.ac.kr Abstract ocument clusterng s an aggregaton of

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

Improving Web Search Results Using Affinity Graph

Improving Web Search Results Using Affinity Graph Improvng Web Search Results Usng Affnty Graph Benyu Zhang, Hua L 2, Y Lu 3, Le J 4, Wens X 5, Weguo Fan 5, Zheng Chen, We-Yng Ma Mcrosoft Research Asa, 49 Zhchun Road, Bejng, 00080, P. R. Chna {byzhang,

More information

Background Removal in Image indexing and Retrieval

Background Removal in Image indexing and Retrieval Background Removal n Image ndexng and Retreval Y Lu and Hong Guo Department of Electrcal and Computer Engneerng The Unversty of Mchgan-Dearborn Dearborn Mchgan 4818-1491, U.S.A. Voce: 313-593-508, Fax:

More information

Classic Term Weighting Technique for Mining Web Content Outliers

Classic Term Weighting Technique for Mining Web Content Outliers Internatonal Conference on Computatonal Technques and Artfcal Intellgence (ICCTAI'2012) Penang, Malaysa Classc Term Weghtng Technque for Mnng Web Content Outlers W.R. Wan Zulkfel, N. Mustapha, and A. Mustapha

More information

Domain Thesaurus Construction from Wikipedia *

Domain Thesaurus Construction from Wikipedia * Internatonal Conference on Computer, Networks and Communcaton Engneerng (ICCNCE 2013) Doman Thesaurus Constructon from Wkpeda * WenKe Yn 1, Mng Zhu 2, TanHao Chen 2 1 Department of Electronc Engneerng

More information

Reducing Frame Rate for Object Tracking

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

More information

CMPS 10 Introduction to Computer Science Lecture Notes

CMPS 10 Introduction to Computer Science Lecture Notes CPS 0 Introducton to Computer Scence Lecture Notes Chapter : Algorthm Desgn How should we present algorthms? Natural languages lke Englsh, Spansh, or French whch are rch n nterpretaton and meanng are not

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

Fast Computation of Shortest Path for Visiting Segments in the Plane

Fast Computation of Shortest Path for Visiting Segments in the Plane Send Orders for Reprnts to reprnts@benthamscence.ae 4 The Open Cybernetcs & Systemcs Journal, 04, 8, 4-9 Open Access Fast Computaton of Shortest Path for Vstng Segments n the Plane Ljuan Wang,, Bo Jang

More information

TN348: Openlab Module - Colocalization

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

More information

The Shortest Path of Touring Lines given in the Plane

The Shortest Path of Touring Lines given in the Plane Send Orders for Reprnts to reprnts@benthamscence.ae 262 The Open Cybernetcs & Systemcs Journal, 2015, 9, 262-267 The Shortest Path of Tourng Lnes gven n the Plane Open Access Ljuan Wang 1,2, Dandan He

More information

Cross-Language Information Retrieval

Cross-Language Information Retrieval Feature Artcle: Cross-Language Informaton Retreval 19 Cross-Language Informaton Retreval Jan-Yun Ne 1 Abstract A research group n Unversty of Montreal has worked on the problem of cross-language nformaton

More information

Recommendations of Personal Web Pages Based on User Navigational Patterns

Recommendations of Personal Web Pages Based on User Navigational Patterns nternatonal Journal of Machne Learnng and Computng, Vol. 4, No. 4, August 2014 Recommendatons of Personal Web Pages Based on User Navgatonal Patterns Yn-Fu Huang and Ja-ang Jhang Abstract n ths paper,

More information

Scheduling Remote Access to Scientific Instruments in Cyberinfrastructure for Education and Research

Scheduling Remote Access to Scientific Instruments in Cyberinfrastructure for Education and Research Schedulng Remote Access to Scentfc Instruments n Cybernfrastructure for Educaton and Research Je Yn 1, Junwe Cao 2,3,*, Yuexuan Wang 4, Lanchen Lu 1,3 and Cheng Wu 1,3 1 Natonal CIMS Engneerng and Research

More 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

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

11. HARMS How To: CSV Import

11. HARMS How To: CSV Import and Rsk System 11. How To: CSV Import Preparng the spreadsheet for CSV Import Refer to the spreadsheet template to ad algnng spreadsheet columns wth Data Felds. The spreadsheet s shown n the Appendx, an

More information

Interfaces for networked media exploration and collaborative annotation

Interfaces for networked media exploration and collaborative annotation Interfaces for networked meda exploraton and collaboratve annotaton Preetha Appan Bageshree Shevade Har Sundaram Davd Brchfeld Arts Meda and Engneerng Program, AME-TR-2004-11 Arzona State Unversty Tempe,

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

Adaptive Knowledge-Based Visualization for Accessing Educational Examples

Adaptive Knowledge-Based Visualization for Accessing Educational Examples Adaptve Knowledge-Based Vsualzaton for Accessng Educatonal Examples Peter Bruslovsky, Jae-wook Ahn, Tbor Dumtru, Mchael Yudelson School of Informaton Scences, Unversty of Pttsburgh {peterb, jaa38, mvy3}@ptt.edu

More information

Video Proxy System for a Large-scale VOD System (DINA)

Video Proxy System for a Large-scale VOD System (DINA) Vdeo Proxy System for a Large-scale VOD System (DINA) KWUN-CHUNG CHAN #, KWOK-WAI CHEUNG *# #Department of Informaton Engneerng *Centre of Innovaton and Technology The Chnese Unversty of Hong Kong SHATIN,

More information

Arabic Text Classification Using N-Gram Frequency Statistics A Comparative Study

Arabic Text Classification Using N-Gram Frequency Statistics A Comparative Study Arabc Text Classfcaton Usng N-Gram Frequency Statstcs A Comparatve Study Lala Khresat Dept. of Computer Scence, Math and Physcs Farlegh Dcknson Unversty 285 Madson Ave, Madson NJ 07940 Khresat@fdu.edu

More information

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

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

More information

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

Cross-lingual Pseudo Relevance Feedback Based on Weak Relevant Topic Alignment

Cross-lingual Pseudo Relevance Feedback Based on Weak Relevant Topic Alignment Cross-lngual Pseudo Relevance Feedback Based on Weak Relevant opc Algnment WANG Xu-wen Insttute of Medcal Informaton & Lbrary, Chnese Academy of Medcal Scences, Beng 100020 wang.xuwen@mcams.ac.cn ZHANG

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

A new query expansion method based on query logs mining1

A new query expansion method based on query logs mining1 Internatonal Journal on Asan Language Processng, 19 (1): 1-12 1 A new query expanson method based on query logs mnng1 Zhu Kunpeng, Wang Xaolong, Lu Yuanchao School of Computer Scence and Technology, Harbn

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

Ranking Techniques for Cluster Based Search Results in a Textual Knowledge-base

Ranking Techniques for Cluster Based Search Results in a Textual Knowledge-base Rankng Technques for Cluster Based Search Results n a Textual Knowledge-base Shefal Sharma Fetch Technologes, Inc 841 Apollo St, El Segundo, CA 90254 +1 (310) 414-9849 ssharma@fetch.com Sofus A. Macskassy

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

A Clustering Algorithm for Chinese Adjectives and Nouns 1

A Clustering Algorithm for Chinese Adjectives and Nouns 1 Clusterng lgorthm for Chnese dectves and ouns Yang Wen, Chunfa Yuan, Changnng Huang 2 State Key aboratory of Intellgent Technology and System Deptartment of Computer Scence & Technology, Tsnghua Unversty,

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

CS47300: Web Information Search and Management

CS47300: Web Information Search and Management CS47300: Web Informaton Search and Management Prof. Chrs Clfton 15 September 2017 Materal adapted from course created by Dr. Luo S, now leadng Albaba research group Retreval Models Informaton Need Representaton

More information

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

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

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

The Codesign Challenge

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

More information

Federated Search of Text-Based Digital Libraries in Hierarchical Peer-to-Peer Networks

Federated Search of Text-Based Digital Libraries in Hierarchical Peer-to-Peer Networks Federated Search of Text-Based Dgtal Lbrares n Herarchcal Peer-to-Peer Networks Je Lu School of Computer Scence Carnege Mellon Unversty Pttsburgh, PA 15213 jelu@cs.cmu.edu Jame Callan School of Computer

More information

Federated Search of Text Search Engines in Uncooperative Environments

Federated Search of Text Search Engines in Uncooperative Environments 1 Federated Search of Text Search Engnes n Uncooperatve Envronments Luo S Thess Proposal Language Technology Insttute School of Computer Scence Carnege Mellon Unversty ls@cs.cmu.edu Thess Commttee: Jame

More information

Relevance Feedback Document Retrieval using Non-Relevant Documents

Relevance Feedback Document Retrieval using Non-Relevant Documents Relevance Feedback Document Retreval usng Non-Relevant Documents TAKASHI ONODA, HIROSHI MURATA and SEIJI YAMADA Ths paper reports a new document retreval method usng non-relevant documents. From a large

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

Oracle Database: SQL and PL/SQL Fundamentals Certification Course

Oracle Database: SQL and PL/SQL Fundamentals Certification Course Oracle Database: SQL and PL/SQL Fundamentals Certfcaton Course 1 Duraton: 5 Days (30 hours) What you wll learn: Ths Oracle Database: SQL and PL/SQL Fundamentals tranng delvers the fundamentals of SQL and

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 Method of Hot Topic Detection in Blogs Using N-gram Model

A Method of Hot Topic Detection in Blogs Using N-gram Model 84 JOURNAL OF SOFTWARE, VOL. 8, NO., JANUARY 203 A Method of Hot Topc Detecton n Blogs Usng N-gram Model Xaodong Wang College of Computer and Informaton Technology, Henan Normal Unversty, Xnxang, Chna

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

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

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

Learning to Classify Documents with Only a Small Positive Training Set

Learning to Classify Documents with Only a Small Positive Training Set Learnng to Classfy Documents wth Only a Small Postve Tranng Set Xao-L L 1, Bng Lu 2, and See-Kong Ng 1 1 Insttute for Infocomm Research, Heng Mu Keng Terrace, 119613, Sngapore 2 Department of Computer

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

Information Filtering Using the Dynamics of the User Profile

Information Filtering Using the Dynamics of the User Profile Informaton Usng the Dynamcs of the User Profle Costn Barbu, Marn Smna Electrcal Engneerng and Computer Scence Department Tulane Unversty New Orleans, LA, 70130 {barbu, smna}@eecs.tulane.edu Abstract Ths

More information

A Knowledge Management System for Organizing MEDLINE Database

A Knowledge Management System for Organizing MEDLINE Database A Knowledge Management System for Organzng MEDLINE Database Hyunk Km, Su-Shng Chen Computer and Informaton Scence Engneerng Department, Unversty of Florda, Ganesvlle, Florda 32611, USA Wth the exploson

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

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

An Efficient Garbage Collection for Flash Memory-Based Virtual Memory Systems

An Efficient Garbage Collection for Flash Memory-Based Virtual Memory Systems S. J and D. Shn: An Effcent Garbage Collecton for Flash Memory-Based Vrtual Memory Systems 2355 An Effcent Garbage Collecton for Flash Memory-Based Vrtual Memory Systems Seunggu J and Dongkun Shn, Member,

More information

Impact of Contextual Information for Hypertext Documents Retrieval

Impact of Contextual Information for Hypertext Documents Retrieval Impact of Contextual Informaton for Hypertext ocuments Retreval Idr Chbane and Bch-Lên oan SUPELEC Computer Scence dpt. Plateau de Moulon 3 rue Jolot Cure 9 92 Gf/Yvette France {Idr.Chbane Bch-Len.oan}@supelec.fr

More information

An Indian Journal FULL PAPER ABSTRACT KEYWORDS. Trade Science Inc.

An Indian Journal FULL PAPER ABSTRACT KEYWORDS. Trade Science Inc. [Type text] [Type text] [Type text] ISSN : 97-735 Volume Issue 9 BoTechnology An Indan Journal FULL PAPER BTAIJ, (9), [333-3] Matlab mult-dmensonal model-based - 3 Chnese football assocaton super league

More information

High-Boost Mesh Filtering for 3-D Shape Enhancement

High-Boost Mesh Filtering for 3-D Shape Enhancement Hgh-Boost Mesh Flterng for 3-D Shape Enhancement Hrokazu Yagou Λ Alexander Belyaev y Damng We z Λ y z ; ; Shape Modelng Laboratory, Unversty of Azu, Azu-Wakamatsu 965-8580 Japan y Computer Graphcs Group,

More information

Fast Feature Value Searching for Face Detection

Fast Feature Value Searching for Face Detection Vol., No. 2 Computer and Informaton Scence Fast Feature Value Searchng for Face Detecton Yunyang Yan Department of Computer Engneerng Huayn Insttute of Technology Hua an 22300, Chna E-mal: areyyyke@63.com

More information