چكيده به لاتين
Co-reference resolution is a key task in discourse analysis and in many natural language processing applications such as question answering, summarization, machine translation, information extraction, given that the performance of many other tasks depends on the correct output of this type of system. Pronoun resolution is a major and challenging subpart of co-reference resolution, in which only the resolution of pronouns is considered. The existing co-reference resolution approaches can be classified into two broad categories: linguistic and machine learning approaches. Linguistic approaches need a lot of linguistic information for the resolution process. Acquisition of such information is an error-prone and time-consuming process. In contrast, learning approaches need less linguistic information and provide the state of the art results. In this thesis, using the PCAC-2008 corpus, we present a framework for the use of machine learning methods for Persian pronoun resolution. In this framework we introduce a new architecture for extracting training instance. Then, by using dependency parser, we add some new features for co-reference system and investigate the effect of these features to improve system efficiency. The results show that proposed system improve the F-measure of Persian pronoun resolution system by 11.2 percent.
