R. Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC effect. Computer Networks, 33(1--6):387--401, 2000.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....Guided by some experimental queries, we propose some formal criteria for evaluating and comparing link analysis algorithms. Keywords: link analysis, web searching, hubs, authorities, SALSA, Kleinberg s algorithm, threshold, Bayesian. 1 Introduction In recent years, a number of papers [7, 14, 5, 20, 15, 8] have considered the use of hypertext links to determine the value of different web pages. In particular, these papers consider the extent to which hypertext links between World Wide Web documents can be used to determine the relative authority values of these documents for various search ....

....It is easy to construct simple examples which show that secondary eigenvectors sometimes are, but sometimes are not, indicative of sub communities. We present a few indicative such examples in section 5. 2. 3 The SALSA Algorithm An alternative algorithm, SALSA, was proposed by Lempel and Moran [15]. Like Kleinberg s algorithm, SALSA starts with a similarly constructed Base Set. It then performs a random walk by alternately (a) going uniformly to one of the pages which links to the current page, and (b) going uniformly to one of the pages linked to by the current page. The authority weights ....

[Article contains additional citation context not shown here]

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC effect. In Proceedings of the 9th International World Wide Web Conference, May 2000.

....= BF and h i = F . Thus, the authority weight assigned by the HITS algorithm to node i after n iterations is proportional to the number of (BF ) paths of length n that reach node i. 2.2. 4 The SALSA Algorithm An alternative algorithm, SALSA, was proposed by Lempel and Moran [34], that combines ideas from both HITS and PAGER ANK. As in the case of HITS, visualize the graph G as a bipartite graph, where hubs point to authorities. The SALSA algorithm performs a random walk on the bipartite hubs and authorities graph, alternating between the hub and authority sides. The ....

....algorithm does not really have the same mutually reinforcing structure that Kleinberg s algorithm does. Indeed, a i = A A B(i) the relative authority of site i within a connected component is determined from local links, not from the structure of the component. Lempel and Moran [34] define a similar Markov Chain for the hubs and a hub weight vector that is the stationary distribution of the corresponding random walk. The stationary distribution h is the left eigenvector of the matrix W r W c . Figure 2: A bad example for HITS algorithm 2.2.5 Other ....

[Article contains additional citation context not shown here]

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC effect. In Proceedings of the 9th International World Wide Web Conference, May 2000.

....fully Bayesian approach, but nd that it is too sensitive to prior information to be useful at this point. Keywords: tightly knit communities, link analysis, web searching, hubs, authorities, SALSA, Kleinberg s algorithm, clusters, Bayesian. 1 Introduction In recent years, a number of papers [2, 5, 7, 6, 3, 1] have considered the use of hypertext links to determine the authority value of di erent web pages. In particular, these papers consider the extent to which hypertext links between World Wide Web documents can be used to determine the relative authority values of these documents for various web ....

....papers consider the extent to which hypertext links between World Wide Web documents can be used to determine the relative authority values of these documents for various web searches. Such use of the Web s link structure poses a number of problems. One is the tightly knit community (TKC) e ect [6], in which a number of web pages all link to each other and thus all receive high authority values, even though their true value may be much smaller. In this paper, we present a new algorithm to deal with the TKC e ect, and avoid giving too much weight to pages simply because they are members of ....

[Article contains additional citation context not shown here]

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC eect. In 9th International World Wide Web Conference, May 2000. Full version published in Computer Networks 33 (2000), 387-401.

....web pages. Independently, Brin and Page proposed the PageRank algorithm [5] which become an integral component of the successful Google search engine [9] These two algorithms spawned the research field of link analysis ranking, and they were followed by a substantial amount of research work [4, 3, 13, 16, 2, 1, 15]. Most of link analysis ranking algorithms start with a set of web pages, interconnected with hypertext links. Given the underlying graph, the algorithms extract the principal eigenvector (s) of a matrix associated with the graph, and rank the pages according to the value of each page in this ....

....by good hubs. In PAGERANK good authorities are pointed to by good authorities. The HITS and PAGERANK algorithms were followed by a substantial number of variations and enhancements. Most of the subsequent work follows a similar algebraic approach, manipulating some matrix related to the web graph [4, 3, 13, 16, 2, 1, 15]. Recently, there were some interesting attempts in applying statistical, and machine learning tools for computing authority weights [4, 6, 11] 3. DYNAMICAL SYSTEMS A dynamical system describes a weight propagation scheme on the nodes of a graph. We construct the Base Set as described by ....

[Article contains additional citation context not shown here]

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC effect. In Proceedings of the 9th International World Wide Web Conference, May 2000.

....important technique is the Pagerank technique underlying the Google search engine [8] which assigns a global importance measure to each page on the web that is proportional to the number and importance of other pages linking to it. A number of other approaches have also been proposed, see, e.g. [6, 13, 22, 24, 25, 33], that perform link based ranking either at query time or as a preprocessing step. Integrating term based and other factors: Despite the large amount of work on link based ranking, there is almost no published work on how to efficiently integrate the techniques into a large search engine. We are ....

....with high loads by modifying query execution [7] though no details have been published. Second, we believe our results are interesting in the context of the ongoing discussion about different approaches to link analysis, in particular the issues of preprocessingbased [22, 31, 33] vs. online [24, 25] techniques, and global [31] vs. topic specific [22] vs. query specific [24, 25, 33] techniques. Our results indicate that a global precomputed ordering is highly desirable from a performance point of view, as it allows optimized index layout and pruning. While it is not clear yet that ....

[Article contains additional citation context not shown here]

R. Lempel and S. Moran. The Stochastic Approach for Link-Structure Analysis (SALSA) and the TKC Effect. In Proc. of the 9th Int. World Wide Web Conference, May 2000.

....of a Web site; 4) Most adaptive topic distillation algorithms based on HITS take the relationship between queries and documents into consideration. However, these algorithms do not work well on mining informative structures in the absent of a target query. Furthermore, as described in [8][18], the link analysis algorithms, e.g. HITS, are vulnerable to the effect of nepotistic clique attack and Tightly Knit Community (TKC) The effects will be more significant for mining informative structures of Web sites 575 due to the huge amount of nepotistic links and cliques in a Web site. ....

....definition of entropy, tuv is deftned as follows: a, 1 E(AN, 6) Equation (6) means that the more information a link carries, the larger the weight of the link. While using mutual reinforcement approach, such as in HITS, to perform the link analysis on pages of a Web site, the TKC effect [18] is more obvious than on the base set of a specific query. This is because that the number of cycle links and nepotistic links dominates the set of links in a Web site. As deftned in [18] a tightly knit community is a small but highly interconnected set of sites. Sites in the TKC will score high ....

[Article contains additional citation context not shown here]

R. Lempel, S. Moran. The Stochastic Approach for Link-Structure Analysis (SALSA) and the TKC effect. In 9 International World Wide Web Conference, Amsterdam, Netherlands, May 2000.

....underlying the Google search engine [11] which assigns a global importance measure to each page on the web based on the number and importance of other pages linking to it. Another basic approach, introduced by the HITS algorithm of Kleinberg [28] and subsequently modified and extended, e.g. in [5, 13, 31, 43], first identifies pages of interest through term based techniques and then performs an analysis of only the graph neighborhood of these pages. The success of such link based ranking techniques has also motivated a large amount of research focusing on the basic structure of the web [12] e# cient ....

.... in 1998 as the basis of the Google ranking scheme [11, 35] Another related approach to link based ranking, proposed at around the same time, is the HITS algorithm of Kleinberg [28] Over the last few years, numerous extensions and variations of these basic approaches have been proposed; see [5, 13, 25, 31, 43] for a few examples. We focus here on the implementation of the basic Pagerank algorithm on massive graphs, and on its topic sensitive extension in [25] The initial papers on Pagerank [11, 35] did not discuss the e#cient implementation of the algorithm in detail. The I O e#cient implementation ....

R. Lempel and S. Moran. The Stochastic Approach for Link-Structure Analysis (SALSA) and the TKC E#ect. In Proc. of the 9th Int. World Wide Web Conference, May 2000.

....clustering came not from the IR literature, but rather from an oftencited motivation for the popular PageRank algorithm [5] Broadly speaking, PageRank is based on the amount of time a random surfer would spend on each page. Random walks on the link structure of the Web are also discussed in [11]. We felt that actual user data would be a better measure of importance than this random walk idea. 3.1 The mixture model In this section, we consider models for both forming query groups as well as determining the most relevant URLs for each group. We construct a statistical mixture model to ....

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC effect. In Proceedings of the Ninth International World Wide Web Conference (WWW9), number 33, pages 387--401, Amsterdam, Netherlands, May 2000.

....4 Department of Computer Science, University of Toronto, Toronto, Ontario, Canada M5S 3G4. E mail: tsap cs.toronto.edu. Web: http: www.cs. toronto.edu tsap 0 Finding Authorities and Hubs From Link Structures on the World Wide Web 1 Introduction In recent years, a number of papers [3, 8, 11, 9, 4] have considered the use of hypertext links to determine the value of different web pages. In particular, these papers consider the extent to which hypertext links between World Wide Web documents can be used to determine the relative authority values of these documents for various search ....

....sometimes are, but sometimes are not, indicative of sub communities. The significance of non principal eigenvectors is an important topic that will be discussed in the full version of this paper. 2. 3 The SALSA Algorithm An alternative algorithm, SALSA, was proposed by Lempel and Moran [9]. Like Kleinberg s algorithm, SALSA starts with a similarly constructed Base Set. It then performs a random walk by alternately (a) going uniformly to one of the pages which links to the current page, and (b) going uniformly to one of the pages linked to by the current page. The authority ....

[Article contains additional citation context not shown here]

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC effect. In 9th International World Wide Web Conference, May 2000.

....Guided by some experimental queries, we propose some formal criteria for evaluating and comparing link analysis algorithms. Keywords: link analysis, web searching, hubs, authorities, SALSA, Kleinberg s algorithm, threshold, Bayesian. 1 Introduction In recent years, a number of papers [7, 14, 5, 19, 15, 8] have considered the use of hypertext links to determine the value of different web pages. In particular, these papers consider the extent to which hypertext links between World Wide Web documents can be used to determine the relative authority values of these documents for various search ....

....It is easy to construct simple examples which show that secondary eigenvectors sometimes are, but sometimes are not, indicative of sub communities. We present a few indicative such examples in section 5. 2. 3 The SALSA Algorithm An alternative algorithm, SALSA, was proposed by Lempel and Moran [15]. Like Kleinberg s algorithm, SALSA starts with a similarly constructed Base Set. It then performs a random walk by alternately (a) going uniformly to one of the pages which links to the current page, and (b) going uniformly to one of the pages linked to by the current page. The authority weights ....

[Article contains additional citation context not shown here]

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC effect. In Proceedings of the 9th International World Wide Web Conference, May 2000.

....not a replacement for traditional methods. A complete system will, and should, use both techniques. 2 Clever system and algorithmic modifications The Clever system builds on the HITS algorithm due to Kleinberg [8] Following Kleinberg s original paper, a number of modifications have been studied [1, 3, 4, 5, 10]. For a full description of the Clever system, see [4] here we give a brief overview and discuss subsequent algorithmic modifications pertinent to the present work. The thesis underlying HITS, and Clever, is that content creation on the Web results naturally in two kinds of valuable pages: ....

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC effect. Proc. 9th WWW, 2000.

....Guided by some experimental queries, we propose some formal criteria for evaluating and comparing link analysis algorithms. Keywords link analysis, web searching, hubs, authorities, SALSA, Kleinberg s algorithm, threshold, Bayesian 1. INTRODUCTION In recent years, a number of papers [3, 8, 11, 9, 4] have considered the use of hypertext links to determine the value of di#erent web pages. In particular, these papers consider the extent to which hypertext links between World Wide Web documents can be used to determine the relative authority values of these documents for various search queries. ....

....of sub communities; we present a few indicative such examples in the full version of this paper. The significance of non principal eigenvectors is an important topic that we intend to pursue further. 2. 3 The SALSA Algorithm An alternative algorithm, SALSA, was proposed by Lempel and Moran [9]. Like Kleinberg s algorithm, SALSA starts with a similarly constructed Base Set. It then performs a random walk by alternately (a) going uniformly to one of the pages which links to the current page, and (b) going uniformly to one of the pages linked to by the current page. The authority weights ....

[Article contains additional citation context not shown here]

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC e#ect. In 9th International World Wide Web Conference, Amsterdam, Netherlands, May 2000.

....4 Department of Computer Science, University of Toronto, Toronto, Ontario, Canada M5S 3G4. E mail: tsap cs.toronto.edu. Web: http: www.cs. toronto.edu tsap 0 Finding Authorities and Hubs From Link Structures on the World Wide Web 1 Introduction In recent years, a number of papers [3, 8, 11, 9, 4] have considered the use of hypertext links to determine the value of di erent web pages. In particular, these papers consider the extent to which hypertext links between World Wide Web documents can be used to determine the relative authority values of these documents for various search queries. ....

....sometimes are, but sometimes are not, indicative of sub communities. The signi cance of non principal eigenvectors is an important topic that will be discussed in the full version of this paper. 2. 3 The SALSA Algorithm An alternative algorithm, SALSA, was proposed by Lempel and Moran [9]. Like Kleinberg s algorithm, SALSA starts with a similarly constructed Base Set. It then performs a random walk by alternately (a) going uniformly to one of the pages which links to the current page, and (b) going uniformly to one of the pages linked to by the current page. The authority ....

[Article contains additional citation context not shown here]

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC eect. In 9th International World Wide Web Conference, May 2000.

....formulation of a model similar to Hubs and Authorities; this formulation allows us to invert the search process, i.e. given the URL of a page, we can find the topics the page is an authority on. For differences between HITS and probabilistic approaches, see the work by Lempel and Moran [16]. The notion of adjusting link weights for HITS was studied by Chakarabarti et al. 5] and Bharat and Henzinger [3] Based on the hub and authority structure of a community, Kumar et al. 14] show that a large number of such communities can be identified from their signatures in the form of ....

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (salsa) and the tkc effect. In Proc. of the WWW9 Conference, pages 387--401, Amsterdam, May 2000. Elsevier Science.

....information stored in its KAB. The textual score is computed by measuring the similarity of the pages to the specific query as well as to the agent s domain. The link topology score is computed using a combination of Kleinberg s mutual reinforcement algorithm [4] and stochastic link analysis [5]. The third role of the KA is local search. The KAB pages, as well as the anchor text associated with the links to the satellite pages, are indexed to form the KAB search index. The satellite pages significantly increase the number of resources accessible from the KAB. Given a query, the inverted ....

....scores that the page receives when applying Kleinberg s Mutual Reinforcement algorithm [4] to the weighted graph at hand. o Stochastic hub authority scores. These are the hub authority scores that the page receives by applying SALSA, the Stochastic Approach for Link Structure Analysis [5] to the weighted graph at hand. These scores are normalized and combined to form the overall link topology score. 2.2.5 Computing the overall score of a Web page The textual score T(s) and link topology score L(s) are combined to yield the overall score of each page in C: S(s) acT(s) 1 ....

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC effect. WWW9/Computer Networks, June, 2000, Vol. 33, No. 1-6, pages 387--401.

....the aggregated text of the pages which currently occupy the core set. 2. Analysis of the link structure of the subgraph # assigns each candidate page with a hub score and an authority score. These scores are computed using a combination of Kleinberg s mutual reinforcement algorithm [9] and SALSA [10], a stochastic link analysis algorithm. Following [5] links are weighted according to both the anchor text which is associated with them (and its similarity to the query) and according to their endpoints membership in the evolving core set. The textual and link analysis scores are combined to ....

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC eect. WWW9 / Computer Networks, 33 (1-6):387-401, June 2000.

No context found.

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC effect. Computer Networks, 33(1--6):387--401, 2000.

No context found.

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC effect. Computer Networks (Amsterdam, Netherlands: 1999.

No context found.

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC effect. In Proceedings of the 9th World Wide Web Conference (WWW), 2000.

No context found.

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC effect. Computer Networks (Amsterdam, Netherlands: 1999.

No context found.

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC effect. In Proceedings of the 9th World Wide Web Conference (WWW), 2000.

No context found.

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC effect. Computer Networks (Amsterdam, Netherlands: 1999.

No context found.

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC e#ect. Computer Networks, 33(1--6):387--401, 2000.

No context found.

Ronny Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC e#ect. In The Ninth International WWW Conference, May 2000.

No context found.

R. Lempel & S. Moran, The Stochastic Approach for Link-Structure Analysis (SALSA) and the TKC Effect. Abridged Version. Department of Computer Science,The Technion, Haifa 32000, Israel.

No context found.

R. Lempel and S. Moran. The stochastic approach for linkstructure analysis (SALSA) and the TKC effect. In Proceedings of the 9th International World Wide Web Conference, pages 387--402, 2000.

No context found.

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC effect. Computer Networks (Amsterdam, Netherlands: 1999.

No context found.

R. Lempel and S. Moran. The stochastic approach for linkstructure analysis (SALSA) and the TKC effect. In Proceedings of the 9th International World Wide Web Conference, pages 387--402, 2000.

No context found.

R. Lempel and S. Moran, \The stochastic approach for link-structure analysis (SALSA) and the TKC eect," Computer Networks (Amsterdam, Netherlands: 1999.

No context found.

R. Lempel and S. Moran. The stochastic approach for link-structure analysis (salsa) and the tkc effect. In Proc. 9th International World Wide Web Conference, 2000.

No context found.

R. Lempel and S. Moran. The Stochastic Approach for Link-Structure Analysis (SALSA) and the TKC E#ect. In Proc. 9th WWW Conference, 2000.

No context found.

Ronny Lempel and S. Moran. The stochastic approach for link-structure analysis (SALSA) and the TKC e#ect. In The Ninth International WWW Conference, May 2000.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC