摘要
针对地理课程自动解题,采用SVM学习算法实现地理试题自动分类。通过采用TF-IDF技术提取地理试题文本中的特征关键词,并选取LIBSVM中的Linear核函数进行训练,以构建用于地理试题分类的预测模型。在所收集的地理试题集上的实验结果表明,在22个试题类别上的单类分类精度达到80%以上,整体分类精度达到了87%。
Aiming to automatic problem solving of geography course,this paper applied the SVM algorithm to the classification of geography problems.And it used the TF-IDF technology to extract feature words in the text of the geography test problems,and then selected the Linear kernel function in LIBSVM to construct the prediction model.The experimental results on the collection of geography problems show that the single classification accuracy on 22 categories is above 80%,and the overall classification accuracy is 87%.
作者
朱刘影
杨思春
Zhu Liuying;Yang Sichun(School of Computer Science&Technology,Anhui University of Technology,Maanshan Anhui 243032,China)
出处
《计算机应用研究》
CSCD
北大核心
2018年第9期2707-2710,共4页
Application Research of Computers
基金
安徽省自然科学基金资助项目(1808085MF178)
安徽省高校自然科学研究重点项目(KJ2016A098)
关键词
自动解题
支持向量机
试题分类
automatic problem solving
support vector machine(SVM)
problem classification