الفهرس 
chapter (1)Only 14 pages are availabe for public view
 |  | from | 107 |  |  |
| | | from | 107 | | |
Abstract
Q/A systems belong to the category of advanced IR tools. They differ from the widely used SEs since a precise answer is returned to the user rather than a list of snippets. Indeed, the use of SE presents a constraint for users as they have to manually filter a long set of returned documents. Most researchers agree that open domain Q/A system consists of three phases: Question Analysis, Passage Retrieval, and Answer Extraction. Question Analysis is the first phase which identifies the focus of the question, classifies the question type, derives the Expected Answer Type (EAT), and reformulates the question into semantically equivalent multiple questions. The second phase is the Passage Retrieval (PR) phase which is the core of any Q/A system. It is responsible for creating a set of candidate ordered paragraphs that supposed to contain the answer(s). The third phase is the Answer Extraction (AE) phase which is responsible for identifying, extracting, and validating the answer(s) from the set of ordered paragraphs passed to it from the PR phase. In the proposed system, we have followed the previously mentioned phases. Question Analysis and classification is the first phase in our system. It defines the type of the question, the EAT, and the collection of queries equivalent to the question. A Java program has been implemented to perform all the processing required by this phase. Our system define the question type and the EAT depending on the interrogative noun of the question (ex. “من ”, “متى ”, “أين ”). Then, the system identifies the main question keywords. In order to do so; the system will remove all the Stop Words. The extracted keywords may be found in the text but in different form. In such case, a QE process can be used in order to overcome the situations where the PR process eliminates relevant passages containing other forms of the question keywords or words related to them. Finally combinations of expanded keywords are formulated to be used in the second phase. The second phase in our system is the PR phase. It is the core of Q/A system. This phase accepts queries from the previous phase and produces top ranked passages related to these queries. We use Google search engine and get the first five snippets related to each query. Then, we used JIRS to improve the rank of Google returned passages by re-ranking them according to the structure base. AE is the final phase in Q/A system which responsible for analyzing the documents or passages returned by the previous phase and identifying answer(s) to the question. The answer type gotten from fist phase is used here to identify the target answer. This thesis contains six chapters in addition to references. Contents of the thesis are organized as follows: Chapter (1): presents a brief introduction of the research study which includes the objectives, the problem description and the methodology of study that has been used to solve this problem. Chapter (2): describes the Arabic language definition and challenges facing researchers in NLP domain and how to overcome these challenges. Chapter (3): presents a background on Arabic Q/A system and the history of Q/A system in English and Arabic languages, then we briefly describe Q/A system components. Chapter (4): describes a detailed explanation of the proposed system and the steps followed to complete the system. Chapter (5): demonstrates and analyzes the experimental results of all the experiments conducted in our proposed system. Chapter (6): presents the conclusions and recommendations, which were derived from this thesis, in addition to the future work for this research.