Research And Implementation On Agent Based Question Answering System In Internet

The remote education based on network gets more and more concentration as novel educational fashion due to the development and spread of computer science and network. Question answering (Q/A) takes an important module in the network education system, and it is regarded as an essential link in educational activities. It combines natural language processing technology with information inspection technology to answer questions described in natural language automatically, concisely and accurately. However, there are some common faults such as limit intelligence, bald interactive methods between teachers and students in the existing Q/A systems.In this paper, the shortcomings above are overcome by constructing a model of intelligent online Q/A system based on Agent using some existing natural language processing technologies. And we get a lot of advantages from agent such as intelligence, independency, interaction, and mobility in this Q/A system. It intends to creating an excellent study environment and enhancing the teaching effect.The contributions of this paper are as follows:1. A model of Agent-based intelligent online Q/A system is constructed, in which each user is regarded as an Agent distributed on nodes in network. This system composes of knowledge base, user model database, dictionary, user information database and various Agents including user Agent, question retrieval Agent, and user model Agent, and all these Agents cooperate with each other. It is capable of answering questions independently and jointly, and improving the dynamic real-time interaction.2. Algorithms are analysed, which are in common use of Chinese word dividing, ambiguity analysis, correlation analysis for questions vs. answers. An improved algorithm of Chinese word dividing and ambiguity processing is presented, which are able to process Chinese more quickly and to analyses ambiguity more efficiently.3. A simple function of communication between Agents is proposed and carried out based on FIPA-ACL language.4. A necessary security mechanism for the online Q/A system is made according to the actual needs of the system and the features of the Agent in it.5. An online Q/A system for college computer culture base course is developed based on the model together with the IBM-Aglet platform.In this paper, the online Q/A system have the advantages of intelligence and coordination. The research on improving the intelligence of remote education in this paper accumulates a lot experience for generally enhancing the production of remote education.