| Today is the age of the Internet.Online examination has been widely concerned at home and abroad.Item bank management technology has also been developed in depth.However,there are some deficiencies in the existing item bank management system.At the same time,the massive item bank in the era of big data may lead to the semantic similarity in intelligent test paper generation,which brings new challenges.Relying on the online examination system,this paper is based on solving the new problem of intelligent test paper generation of massive item bank,and then designs and implements one feasible intelligent item bank management system.There are a lot of researches on intelligent test paper generation in item bank management at home and abroad,but few of them involve the processing of semantic similarity.In order to solve the problem of semantic duplication in the Intelligent Test Paper Generation Method,Genetic Particle Swarm Optimization Algorithm was used in this paper to search multi groups of test papers that conforms to the constraints of the test.Then the idea of density entropy was used to screen the test papers,so that test papers will cover more questions uniformly.After that,analyze the semantic similarity of the test papers that use the TextRank algorithm to extract the key words of test questions,and use knowledge-keywords weighted VSM model to calculate the semantic similarity of test questions.Eliminate the high repetition rate of the test papers and avoid the repetition of the inspection points.The algorithm can solve the semantic similarity problem in intelligent test paper generation of mass item bank.Firstly,this paper investigates the resear-ch status of intelligent item bank management system for online examination,and introduces the technology and ideas used in the system.Secondly,by analyzing the user requirements of the system,the core functional indicators and performance requirements of the system are defined.Then,the key problems of semantic repetition that may exist in the system are studied,and an intelligent test paper generation algorithm is proposed to solve the semantic similarity of test questions.Then the general design of the whole system is described in detail,which focuses on several key issues in the design and implementation of the system,and gives effective solutions.Then,the implementation process of each functional module of the system is described completely according to the requirements of the general design,and the design and implementation details of the paper-forming subsystem are introduced in depth,which highlights the intelligence of the system.Finally,the system is tested,and the learning experience of author and the development prospects of the system are summarized. |
