Research And Implementation Of IR-tree Hybrid Index Model Based Mobile Web Search

With the rapid development of mobile Internet, the number of mobile users has increased dramatically, so far mobile phone users accounted for69.3%of the Internet users. In daily life, mobile users often encounter "Now&Where" related issues. Therefore, mobile Web search becomes one of the most popular activities when users access to the network. Mobile Web search and traditional Web search are quite different, mobile Web search most demand are closely related with the position, while the traditional Web search only consider text retrieval, so that it always cannot get satisfaction of the search results in mobile Web search. Therefore, mobile Web search need considering text similarity and position of similar degree.Text-space hybrid index is one of important research direction in mobile Web search field. This paper mainly focuses on a specific model of the text-space hybrid index:IR-tree hybrid index model. IR-tree hybrid index model consists of inverted index and R-tree index, inverted index is used to calculate the text similarity, R-tree index used to calculate position close degree. IR-tree index will be two index structure organic unifies in together to realize text similarity and position the fusion of correlation. I through extensive reading of foreign literatures and materials, researching and analyzing of the IR-tree index model of the principle, construction steps, analyzing the IR-tree model is how to text index and space position index organic unifies in together, studying IR-tree model related series of algorithm, improve insertion algorithm. Finally, I will realize IR-tree model, combine the north Chinese Web test set (CWT) to do a lot of experiments, to prove the correctness of the IR-tree model and compare IR-tree model before and after optimization efficiency.