利用文本相似度改进遗传算法的组卷实现与研究  

Research and implementation of test-paper combination with genetic algorithm improved by text similarity

作　　者：王慧敏[1] 

机构地区：[1]黄淮学院数学科学系,河南驻马店463000

出　　处：《现代电子技术》2016年第5期117-120,125,共5页Modern Electronics Technique

摘　　要：在分析遗传算法适应度对遗传算法收敛性影响的基础上,针对试题试卷部分的信息参数和评价指标,对遗传算法在组卷中的应用进行了具体的论述,分析了遗传算法在组卷系统中各个部分的实现方法,提出了组卷模块的体系结构。针对遗传算法本身的原理特性进行讨论,研究其理论基础及其所具有的过早收敛的"早熟"现象,并利用排挤算法的早熟现象将组卷系统中的遗传算法进行改进。随后建立向量模型进行快速的试题相似性检测,使考试内容和考试质量得以提升。最终将研究结果进行实验对比与分析,并对研究结果进行了工程实现。On the basis of analysis of the influence of genetic algorithm fitness on its convergence, the application of genetic algorithm in test-paper combination is discussed detailed for the information parameter and evaluation index of the examina- tion paper. The implementation methods of using genetic algorithm to each part of test-paper combination system are analyzed to propose the system architecture of test-paper combination module. The theory characteristics of the genetic algorithm itself are discussed to study its theoretical foundation and ＂premature convergence＂ phenomenon. The genetic algorithm in test-paper combination system is improved with premature convergence phenomenon of the crowding algorithm. After that the vector model was established to fast detect the test questions similarity, and promote the examination content and examination quality. The re- search results were compared and analyzed by experiments, and the research results were realized in engineering.

关 键 词：组卷 遗传算法 相似性检索 小生境算法 对比实验 

分 类 号：TN911-34[电子电信—通信与信息系统] TP311.52[电子电信—信息与通信工程]