Information Extraction From Amharic Language Text Knowledge-poor Approach

During the last two decades with the accelerated Internet development a great amount of datarnhave been being accumulated and stored on the Web. We are drowns with much data at office,rnhome either in printable or electronic form. Then finding the relevant information from this massrndata is critical. At this end, information extraction is a technology which creates the structuredrnrepresentation of unstructured texts by extracting relevant entities from them, thereby, makingrnthe data analysis realizable.rnThis work focuses on developing information extraction system from Amharic language text.rnThe proposed system developed using GATE (General Architecture for Text Engineering) textrnprocessing environment using knowledge-poor approach on infrastructure domain. Byrnknowledge-poor approach we mean we are using simple rules and gazetteer list for entityrnidentification. Our proposed Amharic text information extractor consists of three phase’srnnamely preprocessing, extraction and post processing. The preprocessing phases used forrnhandling language specific issues and setting the environment ready for extraction process. Thernsecond phase is the main unit in our model. It basically performs named entity recognition,rncoreference resolution and relation extraction and extract relevant text. The post processing steprnannotates the selected data and presents the extracted information in a structured form.rnVarious evaluation techniques, which are used to evaluate the performance of our proposedrnmodel were used. The usual precision, recall and F-measure were used to measure the efficiencyrnof the proposed work. We have used 24760 instances for training and testing our model. Ourrnevaluation was conducted on name entity recognition component separately and the overallrnsystem as information extraction component. Accordingly, the system achieves the F-measure ofrn89.1 % on the named entity recognition and in the overall it achieves the F-measure of 89.8%.rnKey words: Information Extraction, Amharic Text Information Extraction, CoreferencernResolution, Relation Extraction, GATE

Get Full Work

Report copyright infringement or plagiarism