Research On Semantic Web-oriented Natural Language Query Interface

Currently, the data on the Internet has explosive growth, the vast amounts ofdata has brought unprecedented challenges to the data storage and access toinformation. The semantic web is considered to be the next generation Internet wayof storage. The data which is described in the form of the semantic web not only canbe understood by human, but alse can be understood by the computer. Data in thesemantic web is linked to each other, and linked by semantic tags, forming a hugesemantic network. Now, more and more data released in the form of the semanticweb, but for end-users, the difficulty of obtaining information from the semanticweb is far more than the current search engine. The semantic web is currently onlysupports formal query, and only familiar with the semantic web ontology ofprofessionals will be able to write the correct formal query. For end-users, the formof natural language query is most convenient, easy to acceptable way of query.Natural language query interface as a bridge connects the natural language queryand the formal query of semantic web, has the important research value andpractical value.How to effectively translate the user’s natural query question into formal querylanguage of semantic web is the main research content of this article. This paperresearched English and Chinese two languages on the natural language queryinterface task. The main contributions of this paper are as follows:First, after a full analysis the form of query and found that the form of problemwith the rules of chain, namely the matched resource in the query only hasrelationship between the resource before it and after it. According to this hypothesis,we propose a resources complement based search method of natural language queryinterface. The method achieved good results in the form of fixed, simple grammarproblems.Secondly, after the analysis of experimental results of the resourcecomplementary search method, we find that the resources complement searchalgorithms cannot handle complicated grammar, and do not consider the semanticinformation in the query. Therefore, we propose a query interface algorithm basedon natural language dependency graph, to make full use of semantic information inthe query to guide generation of the formal query. The performance of the method inAmerican geographic data sets is better than the result of resources complementbased search method. In order to test the natural language dependency graph queryinterface algorithm for different language environment based on the ability to adapt, we conducted experiments in Chinese data set. The analysis of results, the effect ofdependency graph algorithm in English was better than that on Chinese.Experiments show that the method based on dependency graphs, in theconversion from natural language queries to formal queries of Semantic Web tasks,at the same time, proved that the natural language query interface can effectivelycomplete the task of Semantic Web information search.