Research And Application Of Intelligent Assessment System For Online Examinations Based On Flex And Fuzzy Theory

With the development and progress of education and educational ideology, more and more disadvantage of the traditional examinations has been exposed, now paperless online examination system has been widely used, but there are still many problems, especially in the field of intelligent marking. Currently, automatic marking technology for objective questions such as single or multiple choice questions, fill in the blank question and true or false questions is very mature and widely used in large-scale examination system. However, as related to theories and technology issues such as artificial intelligence, natural language processsing, pattern recognition, the automatic marking technology for subjective questions such as explanation for the terms, short answers questions and essay questions is far from perfect. Domestic scoring method for subjective questions is still manual marking, but with the increase of students, this will bring much additional workload to the teachers. The computers have high speed, high efficiency and high precision, so they are especially fit for the automatic marking of subjective questions. Therefore, the study of automatic marking of subjective questions using computers is of great practical significance.In this paper, with the characteristics of Chinese language, the Chinese Automatic Segmentation tenology, the Chart algorithm in Syntactic Analysis frield and Single Similar Degree theory in fuzzy mathematics are introduced to the intelligent marking system of subjective questions. Firstly, with the research and analysis of the Maximum Matching algorithm of the Chinese Automatic Segmentation tenology, an improved dictionary mechanism is used, which improves the efficiency and accuracy of Chinese word segmentation. Then the Chart algorithm is improved against the problem that the original algorithm is easy to generate redundant edge and its analysis efficiency is very low. Then through the analysis of the way of thinking when the teachers manually scoring the subjective questions, the concept of the single similar degree is introduced based on the fuzzy mathematics theory and an algorithm of intelligent marking of the subjective questions is designed.The process of the intelligent scoring is divided into three main steps:Chinese word segmentation, syntactic analysis and similarity computation. The word segmentation is implemented using the Maximum Matching algorithm based on the improved dictionary mechanism, the syntactic analysis using the improved Chart algorithm, and the similarity calculation using the single similar degree proximity algorithm. After the analysis of the characteristics and tasks of each step, the design ideology, the description and the implementation process of the algorithms used are given. Then how to build RIA applications based on Flex is systematically explained and an solution to integrate Flex with Hibernate framework and Spring framework is given. Based on this architecture, an intelligent scoring system of subjective questions is implemented, archiving a good scoring effect and having much value in applications.