The XML keyword search has been used widely in the application of XML documents. Most of the XML keyword search approaches are based on the LCA (lowest common ancestor) or its variants, which usually leads to the un-ideal recall and precision. This paper presents a novel XML keyword search method which based on semantic relatives. The method fully considers the semantic characteristics of the XML document structure. Based on the stack, the algorithm is also presented to merge the semantic relative nodes containing the keyword as the results of XML keyword search. The results of experiments have been identified the efficient and efficiency of our method.

Users often have imprecise ideas when searching the autonomous Web databases and thus may not know how to precisely formulate queries that lead to satisfactory answers. This paper proposes a novel flexible query answering approach that uses query relaxation mechanism to present relevant answers to the users. Based on the user initial query and the data distribution, we first speculate how much the user cares about each attribute and assign a corresponding weight to it. Then, the initial query is relaxed by adding the most similar attribute values into the query criteria range. The relaxation order of attributes specified by the query and the relaxed degree on each specified attribute are varied with the attribute weights. The first attribute to be relaxed is the least important attribute. For the relevant result tuples, they are finally ranked according to their satisfaction to the initial query. The efficiency of our approach is also demonstrated by experimental result.