| Retrieval effectiveness and robustness are two of the most important criteria of text retrieval. Over the past decades, numerous techniques have been introduced to enhance text retrieval performance including those using phrases, passages, general dictionaries such as WordNet, word sense disambiguation, automatic query expansion, pseudo-relevance feedback, and external sources assisted feedback.; This Ph.D. dissertation study focuses on improving the text retrieval effectiveness and robustness by extending existing retrieval model and providing new techniques which include: (1) Designing and implementing a new retrieval model. (2) Utilizing concept in text retrieval. (3) Designing and implementing a highly accurate word sense disambiguation algorithm and incorporating it to our information retrieval system. (4) Expanding queries by using multiple dictionaries such as WordNet and Wikipedia. (5) Employing different pseudo relevance feedback into the retrieval system including local, web-assisted, and Wikipedia-assisted feedback and adopting semantic information to pseudo relevance feedback.; In this Ph.D. study, our design decisions are verified through experiments in the retrieval system. Results are evaluated by standard evaluation metrics: precision, recall, mean average precision (MAP), and geometric mean average precision (GMAP). |
