Anthony Scime

Web Mining: Applications and Techniques

Web mining is moving the World Wide Web toward a more useful environment in which users can quickly and easily find the information they need. Web mining uses document content, hyperlink structure, and usage statistics to assist users in meeting their needed information. This book provides a record of current research and practical applications in Web searching. It includes techniques that will improve the utilization of the Web by the design of Websites, as well as the design and application of search agents. This book presents this research and related applications in a manner that encourages additional work toward improving the reduction of information overflow, which is so common today in Web search results.

A semantic taxonomy-based personalizable meta-search agent

We address the problem of specifying Web searches and retrieving, filtering and rating Web pages so as to improve the relevance and quality of hits, based on the users search intent and preferences. We present a methodology and architecture for an agent-based system, called WebSifter II, that captures the semantics of a users decision-oriented search intent, transforms the semantic query into target queries for existing search engines, and then ranks the resulting page hits according to a user-specified weighted-rating scheme. Users create personalized search taxonomies via our weighted semantic-taxonomy tree. Consulting a Web taxonomy agent such as Wordnet helps refine the terms in the tree. The concepts represented in the tree are then transformed into a collection of queries processed by existing search engines. Each returned page is rated according to user-specified preferences such as semantic relevance, syntactic relevance, categorical match, page popularity and authority/hub rating.

Learning for automatic personalization in a semantic taxonomy-based meta-search agent

Providing most relevant page hits to the user is a major concern in Web search. To accomplish this goal, the user must be allowed to express his intent precisely. Secondly, page hit rating mechanisms should be used that take the user’s intent into account. Finally, a learning mechanism is needed that captures a user’s preferences in his Web search, even when those preferences are changing dynamically. Regarding the first two issues, we propose a semantic taxonomy-based meta-search agent approach that incorporates the user’s taxonomic search intent. It also addresses relevancy improvement issues of the resulting page hits by using user’s search intent and preferences based rating. To provide a learning mechanism, we represent the entire rating mechanism of semantic taxonomy-based meta-search agent approach as a feedforward neural network model and adopt the generalized delta rule as our basic learning scheme by modifying it to conform to our framework. Finally, the entire methodology including this learning mechanism is implemented in an agent-based system, WebSifter II. Empirical results of learning performance are also discussed.

WebSifter: an ontology-based personalizable search agent for the Web

The World Wide Web provides access to a great deal of information on a vast array of subjects. In a typical Web search a vast amount of information is retrieved. The quantity can be overwhelming, and much of the information may be marginally relevant or completely irrelevant to the users request. We present a methodology, architecture, and proof-of-concept prototype for query construction and results analysis that provides the user with a ranking of choices based on the users determination of importance. The user initially designs the query with assistance from the users profile, a thesaurus, and previously constructed queries acting as a taxonomy of the information requirements. After the query has returned its results, decision analytic methods and information source reliability information are used in conjunction with the expanded taxonomy to rank the solution candidates.

A Personalizable Agent for Semantic Taxonomy-Based Web Search

This paper addresses the problem of specifying Web searches and retrieving, filtering, and rating Web pages so as to improve the relevance and quality of hits, based on the user’s search intent and preferences. We present a methodology and architecture for an agent-based system, called WebSifter II, that captures the semantics of a user’s decision-oriented search intent, transforms the semantic query into target queries for existing search engines, and then ranks the resulting page hits according to a user-specified weighted-rating scheme. Users create personalized search taxonomies via our Weighted Semantic-Taxonomy Tree. Consulting a Web taxonomy agent such as WordNet helps refine the terms in the tree. The concepts represented in the tree are then transformed into a collection of queries processed by existing search engines. Each returned page is rated according to user-specified preferences such as semantic relevance, syntactic relevance, categorical match, page popularity and authority/hub rating.

Intelligent Web Search via Personalizable Meta-search Agents

This paper addresses several problems associated with the specification of Web searches, and the retrieval, filtering, and rating of Web pages in order to improve the relevance, precision and quality of search results. A methodology and architecture for an agent-based system, WebSifter is presented, that captures the semantics of a users search intent, transforms the semantic query into target queries for existing search engines, and ranks resulting page hits according to a user-specified, weighted-rating scheme. Users create personalized search taxonomies, in the form of a Weighted Semantic-Taxonomy Tree. Consultation with a Web-based ontology agent refines the terms in the tree with positively- and negatively-related terms. The concepts represented in the tree are then transformed into queries processed by existing search engines. Each returned page is rated according to user-specified preferences such as semantic relevance, syntactic relevance, categorical match, and page popularity. Experimental results indicate that WebSifter improves the precision of Web searches, thereby leading to better information.

Microtargeting and Electorate Segmentation: Data Mining the American National Election Studies

Business marketers widely use data mining for segmenting and targeting markets. To assess data mining for use by political marketers, we mined the 1948 to 2004 American National Elections Studies data file to identify a small number of variables and rules that can be used to predict individual voting behavior, including abstention, with the intent of segmenting the electorate in useful and meaningful ways. The resulting decision tree correctly predicts vote choice with 66 percent accuracy, a success rate that compares favorably with other predictive methods. More importantly, the process provides rules that identify segments of voters based on their predicted vote choice, with the vote choice of some segments predictable with up to 87 percent success. These results suggest that the data mining methodology may increase efficiency for political campaigns, but they also suggest that, from a democratic theory perspective, overall participation may be improved by communicating more effective messages that better inform intended voters and that motivate individuals to vote who otherwise may abstain.

Explaining religious terrorism: A data-mined analysis

What is the relationship between religious liberty and faith-based terrorism? The wider literature on freedom and terrorism has failed to reach a conclusive verdict: some hold that restricting civil liberties is necessary to prevent acts of terrorism; others find that respecting such rights undermines support for terrorist groups, thus making terrorism less likely. This article moves the debate on liberty and terrorism forward by looking specifically at terrorism motivated by a religious imperative and a country’s level of religious liberty—something not attempted in previous studies. Using classification data mining, we test a unique dataset on religious terrorism in order to discover the characteristics that contribute to a country experiencing religiously motivated terrorism. The analysis finds that religious terrorism is indeed a product of a dearth of religious liberty. The study concludes by discussing the implications of these findings for policy-makers.

Globalized computing education: Europe and the United States

As computing makes the world a smaller place there will be an increase in the mobility of information technology workers and companies. The European Union has recognized the need for mobility and is instituting educational reforms to provide recognition of worker qualifications. Within computing there have been a number of model curricula proposed in Europe and in the United States. This analysis of these curriculum models and example computing programs finds that in computing worker mobility is also possible across the Atlantic.

Testing terrorism theory with data mining

This research demonstrates the application of multiple data mining techniques to test theories of the macro-level causes of terrorism. The unique dataset is comprised of terrorist events and measures of social, political and economic contexts in 185 countries worldwide between the years 1970 and 2004. The theories are assessed using the iterative expert data mining (IEDM) methodology with classification mining and then association mining. The resulting 100 rules suggest that the level of democracy in a country is an integral part of the explanation for terrorism. This research shows that a multi-method data mining approach can be used to test competing theories in a discipline by analysing large, comprehensive datasets that capture multiple theories and include large numbers of records.