柯佳伶Koh, Jia-Ling李怡慧Lee, Yi-Hui2019-09-05不公開2019-09-052017http://etds.lib.ntnu.edu.tw/cgi-bin/gs32/gsweb.cgi?o=dstdcdr&s=id=%22G060447007S%22.&%22.id.&http://rportal.lib.ntnu.edu.tw:80/handle/20.500.12235/106443無中文摘要In this thesis, we focus on automatically constructing the relationship between disease and symptoms by online encyclopedia and web search result, including the ranking of the candidate symptoms and the condition of why the symptom is related to that symptom. The contribution of this thesis is as follows (1) Search-Based Approach can extract the Conditional Relationship in good performance (2)Conditional Relationship can help user gain more information(3) We build a medical domain Knowledge Base can be implement in NLP tools.medical domain text miningrelation extractionweb-search datamedical domain text miningrelation extractionweb-search data