Classifying Web Data in Directory Structures

Abstract

Abstract. Web Directories have emerged as an alternative to the Search Engines for locating information on the Web. Typically, Web Directories rely on humans putting in significant time and effort into finding important pages on the Web and categorizing them in the Directory. In this paper, we experimentally study the automatic population of a Web Directory via the use of a subject hierarchy. For our study, we have constructed a subject hierarchy for the top level topics offered in Dmoz, by leveraging ontological content from available lexical resources. We first describe how we built our subject hierarchy. Then, we analytically present how the hierarchy can help in the construction of a Directory. We also introduce a ranking formula for sorting the pages listed in every Directory topic, based on the pages' quality, and we experimentally study the efficiency of our approach against other popular methods for creating Directories.

Search personalization through query and page topical analysis

“…By using a topical hierarchy that is manually constructed our hope is that we can capture in a better way the topical interests of the users. In addition, by enriching such a hierarchy with WordNet concepts and by using them when determining the topics of the pages we can achieve good classification accuracy (Stamou et al 2006). It is likely that under a different classification scheme or topical ontology, our model would perform differently but we believe that it could still be directly applied without the need for any modifications.…”

“…This is also attested in a previous work we have carried out (Stamou et al 2006) where we managed to automatically classify about 300,000 Web pages in the ontology's topics in nearly 6 h (including data cleaning and pre-processing time) without any need for prior training or human involvement.…”

“…The main advantages that our approach exhibits over the existing classification schemes is that it manages to automatically detect the suitable topic(s) of a page without any prior classification knowledge. Moreover, in a recent study (Stamou et al 2006) we found that our classification model managed to classify a large set of pages in their respective 156 topics with an overall accuracy of 70% compared to the accuracy of a Bayesian classifier that delivered for the same set of pages and topics an accuracy of about 66%.…”

Extraction and classification of dense implicit communities in the Web graph

“…the link structure) is becoming important in such studies [Glover et al 2002;Amitay et al 2003]. An interesting mixed classification methodology is in Stamou et al [2006] where the training data is provided implicitly by using hand made directories. These techniques can handle single sites and a small group of pages (in supervised mode) detecting high level functionalities among a set of categories defined at training time.…”

Σχεδιασμός Και Υλοποίηση Ενός Συστήματος Αποκομιδής Ορισμένης Πληροφορίας Από Τον Παγκόσμιο Ιστό, Με Τη Χρήση Σημασιολογικών Δικτύων Λημμάτων

“…Άλλεσ τεχνικζσ βαςίηονται ςτθν γνϊςθ οντολογιϊν εννοιϊν, όπωσ το ςθμαςιολογικό δίκτυο WordNet. τθν εργαςία(Stamou et al, 2006)[141] χρθςιμοποιοφνται λεξικζσ αλυςίδεσ(Barzilay, 1997) [13+ όρων που ςχετίηονται ςθμαςιολογικά και οι οποίεσ μποροφν να χρθςιμοποιθκοφν ωσ αναπαραςτάςεισ των προσ κατθγοριοποίθςθ ιςτοςελίδων.Για να κατθγοριοποιθκεί μία ιςτοςελίδα ςε κάποια κατθγορία ςυγκρίνονται οι όροι τθσ αντίςτοιχθσ λεξικισ αλυςίδασ με τουσ κεματικοφσ όρουσ κάκε υποψιφιασ κατθγορίασ. ε διαφορετικό πνεφμα οι Holden N., & Freitas A.,[197] κατθγοριοποιοφν ιςτοςελίδεσ εξάγοντασ χαρακτθριςτικά από τισ μετά-πλθροφορίεσ keywords18 και description19 μίασ Λςτοςελίδασ.…”