| Summary: | 碩士 === 國立成功大學 === 資訊工程學系碩博士班 === 90 === Due to the rapid advancement and popularization of computer hardware and network equipment, traditional paper operations are replacing by electronic transmission and storage ways gradually. Although large amount of electronic information provides rich resource for analyzing, we could not accomplish these works by hand. In this thesis, we propose an Information Extraction based approach to discover relation in text documents. In our approach, we use the part-of-speech tagger to acquire semantic information in documents and begin with extracting meaningful terms from sentences according to the part-of-speech tags. Then we extract the relation information in documents by using the Relation Template. The extracted information, which is in the form of “Term→Relation Description→Term”, provides more clear comment than the relation degree for the relationships in the dataset. Further more, transforming the relation information among terms to the manner of Concept Maps gives people an easier and better way for browsing the relation information in the dataset.
|
|---|
