| As the continuous development of wireless communication and the popularity of mobile devices, subscribers wish to access information from internet at anytime and anywhere. The emergency of Mobile Search, resolved this problem. However, because of the small terminal screen and keyboard, low computing capability, narrow bandwidth and traffic charges, current Mobile Search engine such as Google can not meet the users'precise, smart, efficient and other requirements. Therefore, on a mobile platform, how quickly and efficiently to retrieve information, become an important and urgent research topic.Based on comprehensive analysis of the characters of Mobile Search, combining vertical search technology in the internet scope, used the Wireless Application Protocol technology (the most representative way in the wireless communication), and put forward a mobile vertical search model, which extended the vertical search technology into mobile scopes. Thus, the mobile users could access information what they needed accurately and efficiently, with strong research value and application value. The main content of this dissertation as follows:Firstly, the relative knowledge of Mobile Search was described, and given a general understanding of Mobile Search technology, to pave way for back system design and development.Next, the topic ranking algorithm PageRank's principle, advantages and advantages was studied. According to analysis, PageRank exits "topics drift", an improvement was proposed, considering the two factors of content based and link based. Improved PageRank was used to determine theme topic and sort search result. Through simulation and analysis, it verified that the I-PageRank could improve the accuracy of reptile theme and the Precision of search result, to some extents, could overcome the shortcoming of algorithm itself.In addition, based on open source search tool Lucene, combining system, several key modules were developed, such as Lucene index module and retrieval module. Furthermore, the sever-side search service was implemented.Finally, on the open source mobile platform Android, using MyEclipse plug-function, a simulated Mobile Vertical Search system was developed. Through system testing, it proved that it could improve the accuracy of query topic, which could overcome some problems of traditional Mobile Search engine, such as search results redundant and overlapped. It will open up a new path for Mobile Search, in this way, mobile users could find the information what they need more quickly and easier. |