Abstract:The query condition of traditional criminal query is a text,and query results are ordered documents lists.However,the way is not conducive to present the relationship of the results and help the police find clues of cases.Therefore,this paper reconstructs counterfeit crime data in the form of information network based on heterogeneous information network information,and constructs a counterfeit money crimes information network.We apply name disambiguation to build relationships of the suspects in counterfeit crime information network,and study relevancy problem of counterfeit money crimes information network using learning to rank methods.In this paper we design and implement the counterfeit money crimes information analysis system.To resolve the query problem of the criminal data,we use the entity as a query term and the network graph as the query results.