习题的关联分析及其向量化表示方法

发布时间：2018-04-22 06:02

  本文选题：题向量 + 题向量化模型　； 参考：《计算机工程与科学》2017年10期

【摘要】：随着互联网+教育的深度融合以及移动终端上电子习题的推广使用,学生的学习过程数据可以被实时获取,充分利用这些过程数据,及时定位学生的知识病灶,开具有针对性的辅导处方,实现知识的按需推送,对于减轻学生的简单重复劳动,提高学习效率将会产生积极影响。试图通过分析在线习题系统的答题数据,发现学生的知识掌握规律,根据错题的伴生状况捕获习题的相关性。为此,构建了题向量化模型,提出了题向量表示的新方法,设计了负采样训练算法,并用程序实现了上述算法。经过实际在线系统的相关数据训练,获得了相应题向量,而后利用题向量的向量运算,可方便查找相同习题、相同知识点习题以及相近知识点习题等,可根据学生错题个案,推断其知识掌握的其他薄弱环节。
[Abstract]:With the deep integration of Internet education and the popularization of electronic exercises on mobile terminals, students' learning process data can be obtained in real time. It will have a positive effect on lightening the students' simple repeated work and improving the study efficiency by issuing the directed guidance prescription and realizing the knowledge push according to the need. By analyzing the answer data of the online exercise system, this paper attempts to find out the rules of students' knowledge mastery, and to capture the correlation of the exercises according to the concomitant condition of the wrong questions. In this paper, a problem vectorization model is constructed, a new method of problem vector representation is proposed, a negative sampling training algorithm is designed, and the above algorithm is implemented by program. Through the relevant data training of the actual online system, the corresponding problem vectors are obtained, and then the vector operation of the problem vectors is used to find the same exercises, the same knowledge points exercises and the similar knowledge points exercises, and so on. Infer other weaknesses in his knowledge.
【作者单位】： 江苏师范大学计算机科学与技术学院;
【基金】：国家自然科学基金(61402207,61272297) 江苏省普通高校研究生科研创新计划(KYLX15_1454)
【分类号】：G434;TP311.13
，

本文编号：1785896