Building Of Intelligent Item Bank System

In traditional teaching model, teachers must spend a lot of time and efforts to collect andmanage questions. Faster and faster, the space of development of network technology, thegrowing rang of the questions on network access to teaching resources. Continue to usetraditional questions management style apparently can’t adapt to the modern way of education.Using an Item Bank System to manage those resources effectively and easily could provide astable basement for examination paper generation automatic by computer. The system couldimprove teacher’s working efficiency and teaching quality.Currently, computer Item Bank System technology is already very mature, but absentfrom the aspect of intelligent like item distinguish and analysis. While in the area of computertechnology, the intelligent recognition is a hot and difficult pot. Their performance is notsatisfactory in aspect of questions automatically identify, questions classified automaticallyand automatic generating test paper.This paper discusses three problems which include test identification, test data storing,test classifying automatically and automatic generating test paper. For different problems, thedifferent improvement methods are respectively put forward. Based on the text paragraphprocessing, the text format and the content of each section is analyzed and identified.Bookmarks in Word is used to identified item segmentation results. Item data is stored bycatalog. Firstly, the test content is stored in a good document assigned, and then the testattribute information is updated to the database. Test questions are classified automatically byusing a fuzzy clustering analysis technology. According to the key database and the keyappeared in the test, the similarity degree among different test item is calculated and thecategory that the current test belongs to is obtained. After a lot of study of summary andanalysis of automatic generating test is done, genetic algorithm commonly used in the testpaper program has been improved. On this basis, by improving the choice of operator strategy,the system can better meet the needs of the user’s test paper, but also improve the convergencerate of the sample groups.Finally, the dissertation introduces the design and implementation of a general intelligentmanagement system for general examination questions base. After extensive experimental testing, it is proved to be the system of high reliability and interface friendliness, and it canimprove work efficiency of the teachers and meet the expectation of the user.