الفهرس 
Table of ContentsOnly 14 pages are availabe for public view
 |  | from | 97 |  |  |
| | | from | 97 | | |
Abstract
Web search engines should understand the kind of information a user is looking for. So there is a need to a small set of words –known as a query– to searching for information. The traditional search methods sometimes not satisfy the users, because search methods often fail to understand user’s query intents. Top-ranked documents provided by traditional web search engines are not able to satisfy different user needs since they cover the same piece of relevant information. With the emergence of advanced Web 2.0 based Rich Internet Applications, This brings a challenge for the Web search solutions to let individual users find the right information as per their preferences, because traditional Web search engines have been built on “one size fits for all” concept. Different users of the Web may have different preferences. This dissertation provided a proposed model based on three studies that support information from various sources query log, browsing history, and social network to extract the user’s interests to understand the user’s behaviour and discover the user’s intent to improve the quality of web search results, while increasing the user’s satisfaction to find what they search for in the top results. The First study analyzed the user’s query logs to classify the related queries, the related intent topic categories, and the related intent types and use this classification to dynamically predict the users’ future queries, its intent topic and its intent type. AOL Search Query Log was taken as an experimental data set. Then evaluation metrics were used to evaluate the prediction results. The main objective of the second study was to provide features that could help users during their web search by categorizing related browsing URLs together. That would be done by identifying intent groups for each URLs category, then identifying intent-segments for each intent group. Through the use of the normalized discounted cumulative gain (NDCG), the experimental results showed the proposed method could improve the performance of the search engine from 0.75 to 0.87. The Third study presented a new proposed method of enabling personalized Web search for users based on their extracted interests and intents from search logs and composite social networks. This study explored various extracted features and intents from previous resources. Then the users’ extracted intents were clustered and used it to re-rank the web search results. Through the use of NDCG and MAP metrics, the implementation and the evaluation of the proposed method were improved the performance of the Web search engines from 0.67 to 0.76 and from 35.88 to 41.22 respectively.