An Internet search with a modern search engine involves iterations of composing a query, submitting the query to the search engine via a text box, reviewing the returned search results list for relevant web pages, and visiting potentially relevant web pages. Often times, one iteration of search steps does not satisfy a search engine user's information need, especially if the user knows few keywords that are relevant to the information need. After a few search iterations, the user's patience and determination levels will be tested.; For the purpose of making information searches on the Internet more efficient and productive, in this thesis a distance measure that takes advantage of past user queries is developed. In presenting the distance measure, the internet search process is described as a dialogue between the user and the search engine and methods are defined in the context of Belief, Desire, and Intention Theory. The distance measure is formulated using topology and the measure calculates distances between queries based on the search results of those queries. The tested hypothesis is that there are correlations between user intentions, user queries, and the search results of queries as reflected in the calculated distances between those search results.; To test the hypothesis, analysis was performed on data from a survey administered to 39 subjects. The survey gathered relevant subject background information, queries related to two different search tasks that were presented to the subjects, and subjects' intentions behind their queries with regards to recall and precision. A total of 390 queries were collected from the subjects, of which 334 queries were unique. The proposed distance measure was applied to the search results of those queries and statistical analysis was performed on the survey data to find correlations between query distances and subject background information, experience, and intentions. The results indicate that query surveys show potential for learning more about search engine user behavior and that the proposed distance measure shows potential as a basis for developing Internet search tools.
展开▼
