Research And Realization Of Web-based Automatic Qustion Answering System

In recent years, modern distance education featuring interactive teaching, flexible schooling system, which integrates technology of computer, network communication, modern education, has been gradually accepted. Teaching and learning of modern distance education are separated in time and space. With the growth of students, individual tutorship according to students' personality and constantly improving quality of network education have become its core and key issues. Network Coaching Answering System is an essential part of distance teaching process and basis of student-oriented personal teaching. It's unpractical to have scare teachers answer questions presented by students.Meanwhile, according to experience of subject teaching and practical application, understanding of courses by students generally has similarities and most of the presented questions are same only with different method of expression, which enable automatic answering a need that searches answers automatically according to users' questions in form of natural language and gives back them accurate information concerning answers. Automatic answering system serves to solve students'problems during study timely which maintains their enthusiasm in study and lifts teachers out of repeatedly answering the similar problems before concentrating on teaching reform and research. Therefore, this paper's proposal has its value in academic meaning and applicationOur automatic answering system is a Q&A system based on particular knowledge base which is constructed against network education, a kind of special application. We adopt tree based structure in knowledge base design to improve the efficiency of search. Two modules with independent functions serve as answering components, namely question process module and information search module. The former one conducts parsing and simple semantic analysis on user's questions before obtaining key words, standard type, etc for further search, which lays sound foundation for enhancement of accuracy of answering The latter one searches knowledge base according to question type to locate identical type answer set; then searches the set according to main key words and corresponding thesaurus set, to obtain to-be-selected answer set related with user's questions. Then specific algorithm is used to compute similarity between user's question and each answer (a short sentence) in to-be-selected answer set and return the short sentence with the biggest similarity value as the final answer.This paper focuses on following aspects in study: classification of questions and process technology, technology of organization and construction of knowledge base as well as technology of information search. In addition, we achieve certain innovation in the research on interrogative phrase classification, knowledge base organization, etc. Experimentation is conducted to show that our system boasts sound accuracy and efficiency in search, satisfying requirement of distance education answering and lives up to our expectation.