ParSemKB: Integrating Text Mining Results from Different Modules Using Extended OWL-DL

Statement Hereby I confirm that this thesis is my own work and that I have documented all sources used. Declaration of Consent Herewith I agree that my thesis will be made available through the library of the 2 Acknowledgement I'm indebted to my supervisor Dr. Erica Melis and Prof. Dr. Siekmann, who recommended me to visit the Parsing and Semantics group in Xerox Research Centre Europe (XRCE), where I started to know the field of semantic extraction of information and applied the knowledge representation technologies I've learned from their lectures. Without their continuous support , engagement and of course their patience with me this work would not have been possible. for their introduction to several state-of-the-art issues in the field of information extraction and for their co-work on applying for the invention disclosure (ID) in XRCE. Without their precious description of the current developed text analyzing engine and our discussion on its extension, the implementation and application of this work would not be accomplished successfully. for their patient proofreading and comments. Their effort has made this thesis on an understandable stand. Further more, I would like to thank Dr. Frederique Segond for her solid support and motivation. The Parsing and Semantics group led by her has provided me with a comfortable working atmosphere. I also appreciate the encouragement from Dr. Christ Dance, the discussion on Machine Learning area with Dr. Hong Chang and Dr. Guillaume Bouchard. In all, I appreciate XRCE for the wonderful research environment and the enjoyable time in Grenoble. Last but not least, I would like to thank my friends, concerning every little progress of me. The most precious thank is given to my mother, who lives alone in my hometown for 7 years since my residence in Germany. Without her mental and financial support, I would never be able to finish my study in Germany. 3 4 Summary In this thesis we are interested in combining the semantic analysis of natural language processing (NLP) beyond the individual sentences. Most of the approaches in NLP have an important limitation: sentence-based analysis. When combining the sentences, there is the challenge of the heterogeneity of representations with different focuses. Besides, we want to enhance the semantic analysis based on results from syntactic analysis modules, in order to " understand " the meaning of text and be able to answer deeper queries beyond keywords-based ones. Motivated by these reasons and …