Current Status Research And Improved Design Of Meta Search Engine

The number of websites has rapidly increased and continues to increase with the development of Internet, while information is digitizing, prevailing and globalizing. An investigation shows that there were 2.8 million servers on which stored about 8 thousand million ffeb pages containing 15TB information on Web even on February, 1992. Gathering, broadcast and utilizing information develop so high that people urgently need an effective retrieval tool to help them find the right information quickly in the infinite data domain.Increase speed and special structure of Web information make some difficulties in its retrieval. Search Engine is becoming an independent tool now after developed from simple Robot Search Software, Single Search Engine to Specialized Search Engine and Meta Search Engine since it came into being to play up to people' s requirements.Single Search Engine which provides some service distributed in different fields still is applied widely because its completion is relatively easy, and was created too much since it first appeared. But it can' t keep up with current staggering development of Web with time passing by because its limited coverage and low efficiency. To change the condition, some of them turn the former method by which gathered a variety of information to another which just point the data of the specialized field. The innovation is very effective for clear or the specialized query. You can see that Specialized Search Engine gets high precision in some field just at expense of large retrieval coverage. Meta Search Engine can gain larger one and avoid repetitive queries to different Single Search Engines by integrating them. In addition, it also puts good foundation to improve high precision because it canenlarge choice scope.However, "Lowest-common-denominator" phenomenon that always reduces advantages of Meta Search Engine is prevailing in them now. The reason is that their target source Search Engines are heterogeneous and hard to integrate effectively. Although somebody proposes standardization for interfaces of all Search Engines and for document construction to solve this problem, they are not being applied extensively for a number of reasons (e. g. a large amount of legacy information, authors unwilling to write articles complying with strict rules, etc). Distributed theory suggests to divide the whole information space into some subspace by some divide operators, the Information Retrieval System in each subspace provides separated or associated service to user. It not only can save resource, but also enlarge retrieval coverage, however, it needs related mechanism to complete protocol transformation and cooperative working which is difficult to realize at present.We will take Meta Search Engine as research object and design improved one on the base of analyzing the current virtues and shortcomings of Single Search Engine, Meta Search Engine and distributed retrieval theory, taking science literature retrieval as an example. We will choose object to integrate meticulously and present realization method in detail. By this way, the difficulties of integration can be conquered at most and the big advantages of Meta Search Engine on interface, dispatching and result showing can be brought into play.We will import new methods and ideas from other fields and propose some original solutions based on current status research of Search Engine in the design. All these methods not only help Meta Search Engine synthesize advantages of other kind of Search Engines, but also make it effective and easy to realize. In the end, we will present futuresearch work facing users' requirements for Search Engine.