The Research Of Ontology-based Automatic Construction Method For Health Knowledge Base

With the popularity of online inquiry platform, there have been accumulated a large number of inquiry data. It is a problem for people to extract more useful information accurately from the data and construct a structured knowledge base for using. Information extraction is a technique for data extraction and it can extract structured information from unstructured and semi-structured text.This paper is committed to the research of automatic construction method of health knowledge base. The research is aimed at collecting the inquiry data automatically and extracting disease symptoms, treatment options and check information from them, and constructing a structured knowledge base. We propose an ontology-based information extraction algorithm and save the result data in a structured way.This paper realizes a directed crawler system in the field of inquiry to collect the data, does some analysis and annotation, creates the inquiry ontology with the concept model and relation model in three layers framework of ontology, and fills the ontology instance with collected data and hownet. This paper also proposes the generation algorithm of extraction rules based on keywords and association rules and the ontology-based information extraction algorithm, determines their extraction sequence and source due to their relationship, and classifies and extracts the sentence according to the ontology instances. Among them, the feature based log likelihood ratio algorithm reduces the influence of high frequency keywords in other concepts, the modified FP-growth algorithm based on keyword’ position property filters out the conflicted keywords and improves the relevance of extraction rules, and determining the order of extraction using the ontology relation model and classify the sentence with ontology instances improves the accuracy of the extraction algorithm. The algorithm proposed in this paper is proved more effective through the contrast test. Finally, we implement an automatic construction system of inquiry knowledge base.