An Algorithm for Query Rewriting with Refined Criteria for Query Materialization in Deep Web

In deep web, the properties and constraint on different interfaces are various. The heterogeneity and autonomy make the query between interfaces can not be translated equally. Zhang et al. build up the query translation on three level: property, prediction, query, and constructs a framework to handle query translation between two random interfaces in the same domain. But their works left to be furthered. Hence, we propose one algorithm with refined criteria to obtain minimal subsuming prediction mapping set based on query "materialization". We also propose a method for query rewriting, which can quantify the query materialization.