Open relationship extraction using conditional random fields

Type Thesis

Rate 0.0 /5

Details

Files

Reviews

Download

Summary

With the growth of information and its ever increasing number of patterns in archiving and storing, the need for systems that can automatically extract useful information is getting more prevalent. Most existing search engines are based on Keywords. The method of structuring and indexing is involved with the existence of the series of Keywords in the content of a Webpage. In other words, keyword searching is used for extracting information for entities which are typically nouns. However, this searching method is incapable of searching for relationships between the entities. This study addresses the limitation of Keyword-based systems and Close-ended systems. In addition, this thesis proposes a system that uses the Conditional Random Field (CRF) model to extract the relationship from Open-ended domains. The system is based on Java and includes three modules: Preprocessing, Training, and Extracting. The results of testing for the proposed system are 89.59% in precision and 53.19% in recall.