Font Size: a A A

The Studies On Auto Assessment Technology Subjective Test Of The Online Examination

Posted on:2014-12-27Degree:MasterType:Thesis
Country:ChinaCandidate:Q WangFull Text:PDF
GTID:2268330425966515Subject:Software engineering
Abstract/Summary:PDF Full Text Request
With the rapid development of computer technology, Auto-Evaluation System withcomputer has been widely used in the examination. The use of Auto-Evaluation not only canreduce the labor force, save social resources but also can guarantee the fairness of theexamination. So far, the technology of objective question automated evaluation is alreadyvery mature,but the technology of subjective question automated evaluation hasn’t achievethe level of application.Based on the analysis of the existing automated evaluation method,We proposed twodifferent methods of evaluating subjective question automatically. One is based on fuzzyconceptual graph and another is based on dependency syntax. In the first method we use fuzzyconceptual graph with weight to express the knowledge of answer and we use and get thescore of the answer by calculating the matching degree between the fuzzy conceptual graph ofstudent’s answer and the fuzzy conceptual graph of standard answer. In the second method weuse words and syntax information of the sentence to express the knowledge and we get thestudent’s score by calculating the semantic similarity of dependency path.We test the methods on two different data and get the performance. We find that the twodifferent methods get better performance in different fields and we can set different weight forthe sub-system in different fields. In the end we implement two different sub-system based onthe fuzzy conceptual graph with weight and dependency syntax and then combine them in acomplete system. Compare with the method based on dynamic programming, we can getmore than21%improve in average deviation and prove that the two methods are better thanthe baseline method.
Keywords/Search Tags:Automated assessment of Subjective Tests, fuzzy conceptual graph, dependencysyntax, Knowledge expression, semantic similarity, natural language process
PDF Full Text Request
Related items