Chakrabarti, S., B. Dom & P. Indyk. Enhanced Hypertext Classification Using Hyper-Links, In Proc. ACM SIGMOD Conference, pp. 307-318, 1998.  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... Raghavan, 1998a; Gibson, Kleinberg, Raghavan, 1998b; Kleinberg, 1998] In addition to finding structural components such as hubs and authorities, hyperlinks can also be used to categorize Web pages. However, exploiting this link information is challenging because it is highly noisy. HyperClass [Chakrabarti, Dom, Indyk, 1998] embodies one approach to this problem, making use of robust statistical models such as Markov using random fields together with a relaxation labeling technique. The methodology of influence weights from citation analysis is similar to a link based search method initially used in the Google search ....

Chakrabarti, S, Dom, B. & Indyk, P. Enhanced Hypertext Classification Using Hyper-Links, ACM SIGMOND International Conference, Management of Data, 1998.

....hyperlinks in all. IBM Almaden Research Center K53 B1, 650 Harry Road, San Jose CA 95120. y Computer Science Department, Brown University, Providence, RI. The network of links in this graph have already led to improved Web search [7, 10, 23, 32] and more accurate topicclassification algorithms [14], and has inspired algorithms for enumerating emergent cyber communities [25] The hyperlinks further represent a fertile source of sociological information. Beyond the intrinsic interest of the topology of the Web graph, measurements of the graph and the behavior of users as they traverse the ....

....the graph are of growing commercial interest. 1. 1 Guided tour of this paper In Section 2 we review several algorithms that have been applied to the Web graph: Kleinberg s HITS method [23] the enumeration of certain bipartite cliques [25] and classification algorithms utilizing hyperlinks [14]. In Section 3 we summarize a number of measurements on large portions of the Web graph. We show that the inand out degrees of nodes follow inverse polynomial distributions [15, 20, 29, 33] and we study the distributions of more complex structures. We present measurements about connected ....

[Article contains additional citation context not shown here]

S. Chakrabarti, B. Dom, and P. Indyk. Enhanced hypertext classification using hyperlinks. Proc. ACM SIGMOD, 1998.

....seven hyperlinks (directed edges) to other pages, making for a total of several billion hyperlinks in all. There are several reasons for studying the Web graph. The structure of this graph has already led to improved Web search [6, 8, 21, 29] more accurate topic classification algorithms [11] and has inspired algorithms for enumerating emergent cyber communities [23] The hyperlinks themselves represent a fecund source of sociological information. Beyond the intrinsic interest of the structure of the Web graph, measurements of the graph and of the behavior of users as they traverse ....

.... work Analysis of the structure of the Web graph has been used to enhance the quality of Web search [5, 6, 8, 9, 21, 29] The topics of pages pointed to by a Web page can be used to improve the accuracy of determining the (unknown) topic of this page in the setting of supervised classification [11]. Statistical analysis of the structure of the academic citation graph has been the subject of much work in the Sociometrics community. As we discuss below, Zipf distributions seem to characterize Web citation frequency. Interestingly, the same distributions have also been observed for citations ....

[Article contains additional citation context not shown here]

S. Chakrabarti and B. Dom and P. Indyk. Enhanced hypertext classification using hyperlinks. Proc. ACM SIGMOD, 1998.

....term based categorizer. It is challenging to exploit this link information, however, since it is highly noisy; indeed, we have found that naive use of terms in the link neighborhood of a document can even degrade accuracy. An approach to this problem is embodied in a system called HyperClass [6], which makes use of robust statistical models such as Markov random fields (MRF s) together with a relaxation labeling technique. Using this approach, it obtains improved categorization accuracy by exploiting link information in the neighborhood around a document. The use of the MRF framework ....

....linked pages are known initially, significant improvement can be obtained using relaxation labeling, wherein the category labels of the linked pages and of the page to be categorized are iteratively adjusted until the most probable configuration of class labels is found. Experiments were performed [6] using pre classified samples from Yahoo and the US Patent Database (www.ibm.com patents) Using HyperClass with hyperlinks cut the patent error rate by half and the Yahoo (Web documents) error rate by two thirds. HyperClass is also used in a focused Web crawler[7] which is designed to search ....

S. Chakrabarti and B. Dom and P. Indyk, Enhanced hypertext classification using hyperlinks. ACM SIGMOD Conference on Management of Data, 1998.

....term based categorizer. It is challenging to exploit this link information, however, since it is highly noisy; indeed, we have found that naive use of terms in the link neighborhood of a document can even degrade accuracy. An approach to this problem is embodied in a system called HyperClass [6], which makes use of robust statistical models such as Markov random fields (MRF s) together with a relaxation labeling technique. Using this approach, it obtains improved categorization accuracy by exploiting link information in the neighborhood around a document. The use of the MRF framework ....

....linked pages are known initially, significant improvement can be obtained using relaxation labeling, wherein the category labels of the linked pages and of the page to be categorized are iteratively adjusted until the most probable configuration of class labels is found. Experiments were performed [6] using pre classified samples from Yahoo and the US Patent Database (www.ibm.com patents) Using HyperClass with hyperlinks cut the patent error rate by half and the Yahoo (Web documents) error rate by two thirds. HyperClass is also used in a focused Web crawler[7] which is designed to search ....

S. Chakrabarti and B. Dom and P. Indyk, Enhanced hypertext classification using hyperlinks. ACM SIGMOD Conference on Management of Data, 1998.

No context found.

Chakrabarti, S., B. Dom & P. Indyk. Enhanced Hypertext Classification Using Hyper-Links, In Proc. ACM SIGMOD Conference, pp. 307-318, 1998.

No context found.

S. Chakrabarti, B. Dom, and P. Indyk. Enhanced hypertext classification using hyperlinks. Proc. ACM SIGMOD, 1998.

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