The research on the method of personalized restricted domain Q&A retrieval based on the linguistic model and the user model

The answer retrieval based on the personalized field Q&A system require system to obtain a text list contains the user's interesting with the user's query. Aimed at the traditional answer retrieval method of field Q&A system can't reflect user's personalization requirement, a personalized restricted domain Q&A retrieval method based on the linguistic model and the user model is proposed. Firstly, the method use MSVDD to construct user model, and deduce the spatial distance between document and user model, on that basis define personalization membership sort function. Secondly, constructing the personalized retrieval model that apply to field Q&A combine the language retrieval model that has smooth attribute. The experimental results show this method can effectively improve user's personalized experience and increase the accuracy of retrieval.