An adaptable Question-Answer System using match of multi-sphere knowledge

For a performance enhancement of the current a QAS(Question-Answer System) and web search, the adaptable semantic decision is important which support expression of natural language and meaning of a input sentence about UIDD(User Information Demand Description) is a difficult situation natural language and adaptable semantic decision in a sentence up to now. For overcome these problems, a various research was attempted that decided for search result about UIDD based on information retrieval model, and a lot of a QAS and search engine like Google applied which was currently proposed, but it is difficult to provide the retrieval results that determine flexibly a similarity of meaning. Therefore, this study, match knowledge which use the semantic inference rule which expressed knowledge in multi-sphere knowledge, and inferenced knowledge combine with n-ary representation-based ontology by weight, then determine various semantic similarity about a vocabulary in UIDD. As decide on semantic similarity, we propose an approach which presented a most adaptable answer that input a question to a QAS. And a proposed approach-based, we implement a QAS which service a prescription regarding a symptom and inform a foot reflective point to user. An experiment result of implement system, determine adaptable to a semantic similarity of a vocabulary regarding input query sentence, and show a stable query result, and a proposal approach demonstrated that we were effective.