The Key Technologies Research Of Personal Information Service Based On Agent

The problems facing the information searching in the Internet nowadays are unorganized, distributed, frequently changing with structure complexity information resources. This directly let the user get overloading information or unsatisfied information. Based on Agent Technology of the latest achievements of Artificial Intelligence, It has been given out a model to obtain the personal satisfied searching information.Through analyzing the search engine technology, technology of personal information service and compared with the latest achievements of Artificial Intelligence, It has been tried to put the Artificial Intelligence such as interest model and individual knowledge finding, multi-agent cooperation in my model. This model is simply realized by a spider program to snatch at web, there the recursion algorithm is used and reorganized by the user's interest. The Agent does the service to the users using push technology relied on their preference.My efforts to make here are: (1)Analyzing and researching the current technologies of search engine, structure of system, the principle of working ,development course and tendency, and their applied actuality; in terms of low service of individuation and intelligence that search engine provides, on the basis of structure of information search engine put forward by professor Shizhongzhi, designed a search engine frame model based on Agent. This model composed of multi agents, divided three tiers from top to bottom: man-machine interactive tier, information management tier and information collection tier. The function of every agent is discussed.(2)Researching and analyzing Chinese words segmentation technology, and improving The Forward Directional Maximum Matching arithmetic (It has been called The Forward Direction Maximum Dynamic Matching), and providing thought of "The Forward Direction Maximum Dynamic Matching" in order to decrease recycling times when matching. As the experiment shows, the rate of Chinese words segmentation is improved obviously.(3)When tackling "various interpretations" in Chinese words segmentation, providing the thought of "tolerance of various interpretations," in other words, integrating the results of "The Forward Directional Maximum Matching" and "The Reverse Directional Maximum Matching ", then handing this result to search engine.(4)Designing programs Spider collects web, and showing the result by the type of JTree in Java. The main idea of spider is a recursion process of depth priority key words searching to sites and breadth priority to pages. It has been given out a method to solve the dead loop and overflow in memory in the process.