A method of search intention recognition in specific domain

The researches on new generation search engine more and more focus on the search intention, which answers the question that why people search, instead of how they search or what they search for. Most existed investigations on search intention analysis considered only the information embedded in the search context; however, the information contained in the search query is more important for analyzing the search intention, especially in specific domain. This paper presents a novel search intention recognition method, which makes full use of the information extracted from both the search context and the search query. Firstly, the characteristics of the two datasets are satisfied independently by two algorithms; and then, the search intention is obtained from those characteristics and datasets through an improved algorithm. In addition, by classifying the keywords of the search query on the basis of the artificial vocabularies, the number of learning sample datum has significantly increased, which makes the algorithm more precisely. Finally, a case is studied to verify the validation and precision of the new method.