Research On Quantization Relation Extraction Method Based On Shallow Analysis

Facing of the massive information resources,during the digitized,networked and big data era,the existing information extraction tools cannot capture or discover the information between the intrinsic values of knowledge effectively.The appearance of information extraction technology makes it possible to transfer text information from unclassified ones to structured ones automatically.On the basis of a lot of corpus analysis and technical methods research,focusing on domestic and international present situation in entity relationship information system,the author demonstrates a Chinese entity relation extraction solution.this solution defined quantifiable entity relationships,developed a framework of extracting quantifiable entity relationship,at the same time used and improved GATE — a mature English information extraction system,combined with ICTCLAS — the mature word segmentation and part of speech tagging tool,achieved high precision rate and recall rate quantifiable entity relationship extraction.This article aims at three key aspects to solve Chinese information extraction,namely,Chinese word segmentation and part of speech tagging problem,definition and building approach of quantifiable entity relationship library,and the systematic implementation scheme of quantified entity relationship library.In addition to the use of ICTCLAS solving the difficulties of Chinese word segmentation,the author proposed a method on entity relationship domain classification and attributes quantization,besides,in order to obtain vast amounts of quantifiable entity relationship sets,composed three types of common rules for the entity relationship model aiming at Chinese entities characteristics.After system implementation,the author utilized the existing domain news set to analysis and narrative the implementation of Chinese entity relationship system,discussed system scalability in multilevel,and verified the practicality and efficiency of the system.Through experiments in this paper,we consider that based on the combination of GATE and ICTCLAS,it is a meaningful attempt to extract quantifiable Chineseentities relationship.It basically solved the problem of the difficulty in locating Chinese entity relationship,mining the Chinese entity relationship the quantitative attributes,and established a good framework for the subsequent Chinese information extraction research.