Research On Chinese Named Entity Semantic Relation Extraction Based On Dependency Tree

Chinese named entity semantic relation extraction is a major research direction in the research of relation extraction. Because the unique characteristics of Chinese text, which is different from English text, makes the slow development of the research on Chinese entity semantic relation extraction. It usually has many long sentences in Chinese text and it usually has many entities in these long sentences, these characteristics makes the research on Chinese entity semantic relation extraction become difficult. The thesis focuses on research on this issue.Firstly, the research background of this topic, the history and development of research on semantic relation extraction is introduced, and the representative approaches of research on Chinese named entity semantic relation extraction are mainly discussed.Secondly, aiming at the problem of inefficiency of existing approaches dealing with entity relation extraction, a novel approach was proposed. This new approach proposes seven heuristic rules to extract relation feature sequence through combining with grammar feature of Chinese text, and applies the semantic sequence kernel function added with pattern weight factor with KNN learning algorithm to fulfill the entity relation extraction task.Finally, based on the research above, CERE-DT(Chinese named Entity semantic Relation Extraction based on Dependency Tree) system is designed and implemented. Moreover, the effective of the system is tested through using benchmarks which are collected on the Internet. The precision, recall and F-score values are analyzed. Analyzing the contrast experiment result of this new approach and traditional approaches, it shows that this new approach achieves better f-score result.