A Retrieval System Based On Virtual Document Of Semantic Association

After mining a large number of semantic associations from massive semantic data, it is a huge challenge to use the mining results. Users are often interested in a certain area or a certain subject of semantic associations, it is necessary to provide the retrieval service. We must solve the short text problem since semantic associations don’t like web pages which contain much information. Before semantic associations returned to user, they need to be ranked in a reasonable way. How to help user to understand the result needs to be discussed, since there are little research about summary of semantic associations.We do some research about retrieval of semantic associations in this paper, the main work is as follows. (1) We apply the index method for semantic associations based on virtual documents. (2) We present a multi-target search model for retrieval of semantic associations. Users can submit different groups of keywords, each group can contain some keywords and will match a specific object. The query result is the intersection of semantic associations that contain all specific objects. (3) We present a ranking method for semantic associations. Two kinds of ranking factors are proposed in this paper:static factors and dynamic factors. Dynamic factors are related to keywords, while the static factors are nothing to do with the keywords. (4) We apply the summarize method for semantic associations. When the system returns the query results, we use the Steiner tree algorithm to calculate the summary by constructing focused association graph, which help users to get a better understand to the results. (5) We give an implement of retrieval system for semantic association.Our work for retrieval and ranking of semantic associations provides a new train of thought and method, which has a certain reference value on research of processing of semantic data.