Research On Technologies Of Intelligent Exam System

With the rapid development of the computer and network technology,the existing teaching content and teaching methods are facing the unprecedented challenges.The intelligent Exam system can be more objective and evaluate the students'true level equitably,furthermore it can economize manpower,material resources,and improve examination efficiency.This thesis makes a research on the technologies of the intelligent exam system.In the text of this thesis,it analyzes all kinds of evaluation factors,the effect of these factors and the relationship between them.On the basis of these knowledge,it adopts the sum of the weighted indicators to construct fitness function of test paper auto-generation algorithm;and the paper analyzes relevant group book tactics,combining with the systematic method merit of auto generating paper in the past,choosing genetic algorithm in being fit to owe systematic algorithm.Aiming at inefficient search and vulnerable emerging immature convergence of the traditional genetic algorithm,this thesis brings forward a new progressive genetic algorithm.The algorithm improves the test paper strategy,coding schemes, fitness function,and the methods of selective and cross mutation operator.In order to further the optimal ability of genetic algorithm's selection operator,a new population-sorted multi-roulette-wheel selection operator is introduced based on traditional roulette-wheel selection operator.Meanwhile,niche technology is adopted,which can keep the diversity of the population and choose the best individual to the next generation directly.The experiments show that the new genetic algorithm has significantly improved in the global search performance and convergence rate.Through practice,it can efficiently examine the knowledge of students.The thesis analyzes the characteristics of nonstandard test models of various courses.It compares the test judging methods of manifold nonstandard test models.Besides,taking the requirement of C Language Programming and functional needs of marking system,via analysis of the artificial thinking habit of going over examination papers,it puts forward an automatic marking model of subjective questions based on keywords match and segmentation similarity.It presents the key technology of this mode:automatic segmentation of subjective answers,automatic match of keywords and automatic marking method of subjective questions based on segmentation similarity.It also carries out to realize this model by parameters adjustment for automatic marking of subjective questions. Using the method mentioned in this thesis to auto-checking programs can obtain a similar result as manual checking.Finally,the engineering implement problems of the automatic generating test paper system based on genetic algorithms are discussed in detail,some codes are provided.The experiments show that this examination system provides more convenient,safer,rational,scientific evaluation of teaching.