Document categorization is one of the foundational problems in (web) information retrieval. Even though web documents are hyperlinked, most proposed classification techniques take little advantage of the link structure and rely primarily on text features, as it is not immediately clear how to make link information intelligible to supervised machine learning algorithms. This paper introduces a link-based approach to classification, which can be used in isolation or in conjunction with text-based classification. Various large-scale experimental results indicate that link-based classification is on par with text-based classification, and the combination of the two offers the best of both worlds.

@techreport{gyongyi2006web,
    author = {Zolt\'{a}n Gy\"{o}ngyi and Hector Garcia-Molina and Jan Pedersen},
    title = {Web Content Categorization Using Link Information},
    institution = {Stanford University},
    year = {2006--2007}}

Answers represents a new type of community portal that allows users to post questions and/or answer questions asked by other members of the community, already featuring a very large number of questions and several million users. Other recently launched services, like Microsoft's Live QnA and Amazon's Askville, follow the same basic interaction model. The popularity and the particular characteristics of this model call for a closer study that can help a deeper understanding of the entities involved, their interactions, and the implications of the model. Such understanding is a crucial step in social and algorithmic research that could yield improvements to various components of the service, for instance, personalizing the interaction with the system based on user interest. In this paper, we perform an analysis of 10 months worth of Yahoo! Answers data that provides insights into user behavior and impact as well as into various aspects of the service and its possible evolution.

@inproceedings{gyongyi2008questioning,
    author = {Zolt\'{a}n Gy\"{o}ngyi and Georgia Koutrika and Jan Pedersen and Hector Garcia-Molina},
    title = {Questioning Yahoo! Answers},
    booktitle = {Proceedings of the First Workshop on Question Answering on the Web},
    location = {Beijing, China},
    year = {2008}}

Tagging systems allow users to interactively annotate a pool of shared resources using descriptive tags. As tagging systems are gaining in popularity, they become more susceptible to tag spam: misleading tags that are generated in order to increase visibility of some resources or simply to confuse users. We introduce a framework for modeling tagging systems and user tagging behavior. We also describe a method for ranking documents matching a tag based on taggers' reliability. Using our framework, we study the behavior of existing approaches under malicious attacks and the impact of a moderator and our ranking method.

@inproceedings{koutrika2007combating,
    author = {Georgia Koutrika and Frans Adije Effendi and Zolt\'{a}n Gy\"{o}ngyi and Paul Heymann and Hector Garcia-Molina},
    title = {Combating Spam in Tagging Systems},
    booktitle = {Proceedings of the 3rd International Workshop on Adversarial Information Retrieval on the Web},
    location = {Banff, Alberta, Canada},
    year = {2007}}

Link spamming intends to mislead search engines and trigger an artificially high link-based ranking of specific target web pages. This paper introduces the concept of spam mass, a measure of the impact of link spamming on a page's ranking. We discuss how to estimate spam mass and how the estimates can help identifying pages that benefit significantly from link spamming. In our experiments on the host-level Yahoo! web graph we use spam mass estimates to successfully identify tens of thousands of instances of heavy-weight link spamming.

@inproceedings{gyongyi2006link,
    author = {Zolt\'{a}n Gy\"{o}ngyi and Pavel Berkhin and Hector Garcia-Molina and Jan Pedersen},
    title = {Link Spam Detection Based on Mass Estimation},
    booktitle = {Proceedings of the 32nd International Conference on Very Large Databases},
    publisher = {ACM},
    location = {Seoul, Korea},
    year = {2006}}

Link spam is used to increase the ranking of certain target web pages by misleading the connectivity-based ranking algorithms in search engines. In this paper we study how web pages can be interconnected in a spam farm in order to optimize rankings. We also study alliances, that is, interconnections of spam farms. Our results identify the optimal structures and quantify the potential gains. In particular, we show that alliances can be synergistic and improve the rankings of all participants. We believe that the insights we gain will be useful in identifying and combating link spam.

@inproceedings{gyongyi2005link,
    author = {Zolt\'{a}n Gy\"{o}ngyi and Hector Garcia-Molina},
    title = {Link Spam Alliances},
    booktitle = {Proceedings of the 31st International Conference on Very Large Databases},
    publisher = {ACM},
    pages = {517--528},
    location = {Trondheim, Norway},
    year = {2005}}

Web spamming refers to actions intended to mislead search engines into ranking some pages higher than they deserve. Recently, the amount of web spam has increased dramatically, leading to a degradation of search results. This paper presents a comprehensive taxonomy of current spamming techniques, which we believe can help in developing appropriate countermeasures.

@inproceedings{gyongyi2005web,
    author = {Zolt\'{a}n Gy\"{o}ngyi and Hector Garcia-Molina},
    title = {Web Spam Taxonomy},
    booktitle = {Proceedings of the First International Workshop on Adversarial Information Retrieval on the Web},
    location = {Chiba, Japan},
    year = {2005}}

Web spam pages use various techniques to achieve higher-than-deserved rankings in a search engine's results. While human experts can identify spam, it is too expensive to manually evaluate a large number of pages. Instead, we propose techniques to semi-automatically separate reputable, good pages from spam. We first select a small set of seed pages to be evaluated by an expert. Once we manually identify the reputable seed pages, we use the link structure of the web to discover other pages that are likely to be good. In this paper we discuss possible ways to implement the seed selection and the discovery of good pages. We present results of experiments run on the World Wide Web indexed by AltaVista and evaluate the performance of our techniques. Our results show that we can effectively filter out spam from a significant fraction of the web, based on a good seed set of less than 200 sites.

@inproceedings{gyongyi2004combating,
    author = {Zolt\'{a}n Gy\"{o}ngyi and Hector Garcia-Molina and Jan Pedersen},
    title = {Combating web spam with {T}rust{R}ank},
    booktitle = {Proceedings of the 30th International Conference on Very Large Databases},
    publisher = {Morgan Kaufmann},
    pages = {576--587},
    location = {Toronto, Ontario, Canada},
    year = {2004}}

E-mail spam is a nuisance that every user has come to expect. But Web spammers prey on unsuspecting users and undermine search engines by subverting search results to increase the visibility of their pages.

@article{gyongyi2005spam,
    author = {Zolt\'{a}n Gy\"{o}ngyi and Hector Garcia-Molina},
    title = {Spam: It's Not Just for Inboxes Anymore},
    journal = {IEEE Computer Magazine},
    volume = {38},
    number = {10},
    pages = {28--34},
    month = {October},
    year = {2005}}