الفهرس | Only 14 pages are availabe for public view |
Abstract This thesis proposes a framework for recognizing, annotating and extracting important entities from unstructured Arabic text and transferring these entities into meaningful knowledge depending on extracted relations between these entities to help decision makers benefit from this knowledge. The proposed solution is applied on the stock market Arabic news text; this is done through creating a concept map that covers a set of developed patterns which represent important domain core concepts such as company’s capital, stocks, profit, loss, income, investment and their relations. These patterns were supported by using text mining tools including R project for statistical computing and frequent patterns analysis project.An incremental algorithm for supporting ontology construction process and updating an existing ontology is proposed. The solution is implemented through Gate for text engineering tool using Gazetteer and Jape rules modules and evaluated on 150 news article text file as 87% error free correct annotation result with performance measures of 100% precision, 87% recall, and F1-score of 93%. This thesis proposes a framework for recognizing, annotating and extracting important entities from unstructured Arabic text and transferring these entities into meaningful knowledge depending on extracted relations between these entities to help decision makers benefit from this knowledge. The proposed solution is applied on the stock market Arabic news text; this is done through creating a concept map that covers a set of developed patterns which represent important domain core concepts such as company’s capital, stocks, profit, loss, income, investment and their relations. These patterns were supported by using text mining tools including R project for statistical computing and frequent patterns analysis project.An incremental algorithm for supporting ontology construction process and updating an existing ontology is proposed. The solution is implemented through Gate for text engineering tool using Gazetteer and Jape rules modules and evaluated on 150 news article text file as 87% error free correct annotation result with performance measures of 100% precision, 87% recall, and F1-score of 93%. This thesis proposes a framework for recognizing, annotating and extracting important entities from unstructured Arabic text and transferring these entities into meaningful knowledge depending on extracted relations between these entities to help decision makers benefit from this knowledge. The proposed solution is applied on the stock market Arabic news text; this is done through creating a concept map that covers a set of developed patterns which represent important domain core concepts such as company’s capital, stocks, profit, loss, income, investment and their relations. These patterns were supported by using text mining tools including R project for statistical computing and frequent patterns analysis project.An incremental algorithm for supporting ontology construction process and updating an existing ontology is proposed. The solution is implemented through Gate for text engineering tool using Gazetteer and Jape rules modules and evaluated on 150 news article text file as 87% error free correct annotation result with performance measures of 100% precision, 87% recall, and F1-score of 93%. |