Design Of Subjective Question Evaluation Algorithm Based On Bidirectional Matching Space Model And Implementation Of Marking System

Subjective review is a crucial function for realizing an online examination system.How to improve the accuracy of subjective correction is one of the key issues.Compared with manual marking,subjective topic automatic scoring technology not only reduces the subjective factors,but also saves the time of marking.How to automatically score subjective questions is also one of the hot spots in natural language processing research.Based on the analysis of existing subjective algorithms both at home and abroad,this dissertation proposes a two-dimensional traversal space model algorithm.The algorithm is designed from two aspects:the key points and the closeness of the answers.(1)Due to the high professional vocabulary requirement during the professional course judgment process,the key points of the student's answer are calculated based on the keyword two-way matching algorithm.(2)When calculating the closeness of the answer,the two answers are regarded as texts,and they are converted into the problem of calculating the similarity between the two texts.The natural word processing technology is used to segment the answers in Chinese,combined with HowNet.The semantic vector spacemodel finds the similarity between words and words,uses the cosine vector algorithm to find the similarity of the words,and finally calculates the final score according to the score weights.(3)Based on the research of bidirectional traversal space model algorithm,a non-line examination system based on ASP.NET MVC framework is developed.The distributed view PartialView and component loading ComponentConfig are used to dynamically generate text boxes/checks using EXT.NET technology.Boxes and other controls for selecting/filling blanks.The experimental results show that the accuracy of the algorithm is high.At present,the system has been actually used and the effect is good.