Question And Answer System Intention Perception Based On The Knowledge Base

Web Intelligence (WI) is gaining its growth in a rapid speed, and theintelligence question and answer system, which returns the answer in aninteractive and understandable way based on the huge web resources, is thecruial stage in the development of Web Intelligence. However, currentquestion and answer system normally applies the recognition of semanticsimilarities in the prepared database, neglecting the true intention hiding inthe expression. In this paper, we present a model based on the medicalquestion and answer knowledge base to overcome this challenge. Theknowledge base includes three parts: disease entity, medicine, and symptomproperties. We extract the entities from the Internet resources and form therelationship between these entities. Then a simple graph path algorithm basedon words detection and relation weight adjustment is used to realize thequestion and answer system intention perception. The experimental resultsshow that our method can effectively perceive the demand of user. Thismethod we proposed can also be applied in deep understanding of otherintelligence systems such as classification and text mining.