Adaptive Web Search Engine for Mobile Devices

Abstract

Web has become the major information source around the world. It grows exponentially by providing more and more information everyday. Hence locating desired information is a critical issue faced by individuals. Web search engines plays as the ideal tool for locate such information effectively and efficiently. And the success of each and every search engines heavily depends on the satisfaction level of individuals. If the search engine is capable of providing relevance and up to date information to the user, it is said to be ultimate tool and would become popular among people. Until recently, web search was done with the assistance of computers and internet technologies. Nowadays, with the increasing popularity of wireless technologies, people access the web through their mobile devices. Mobile search engines have became a novel derivative of searching on the web, which assists mobile users to access the web anywhere in the world. Though mobile search is very popular and convenient approach to access the web, it has some limitations. In contrast to desktop search, using tiny screen mobile devices have reasonable amount of limitations. Interface issues such as small screen, tiny keypad which distracts usersystem interaction, less bandwidth of wireless internet access and high cost are still serious barriers. Though search engine presents loads of results to them, they are interested in specific destinations. They would like to accomplish their information needs by following as few web pages as possible. This will reduce the overhead of transmitting large amount of data to the mobile device and visiting multiple links. But these few pages should be most relevant and effective to their desire. Consequently, providing a well focused search and delivering relevant information is crucial. Our system assumes that mobile device is unique to each individual. Almost all the requests from that device reveal the personal interests of particular user. It further supposes that users usually have long standing interests that will not change drastically over time. As a result our system supposed to maintain a profile for each user which 12 represents personalized interests of each individual. User profile is simply a collection of preferred words of user. This profile can be applied to measure the relevancy of search results. It is known fact that mobile users are not bothering about entering meaningful queries to search engine. Further, they tend to present short queries as well. This occurs mainly due to the tiny device with nine digits keypad which distracts interaction between user and system. In previous study of search queries shown that short queries do not provide enough information to the search engine. To avoid this drawback, our system proposes an approach to refine the user s original query internally in the system. This is done without an awareness of the user. It is expected that reformulated query will be a better representation of individual s information need. As a response, our system proposes new approach to present few most relevant results to the users. This will reduce the amount of data transmitted to the device as well usersystem interaction through minimum navigation across web pages and scroll within the page.