基于文本数据挖掘的硕士论文分类技术

Categorization of master thesis based on text data mining

针对硕士论文的摘要和关键词等数据进行数据挖掘,实现硕士论文自动分类。为此收集了2000余个相关数据,在对所研究的数据对象特点进行分析的基础上,确定了分类算法,详细分析了支持向量机分类方法;对收集的研究数据进行了仿真实验,并与其他常用分类器进行比较。实验表明,基于支持向量机的分类方法比其他常用分类器具有较高的准确率。对实验结果中得到的知识进行了分析,得出一系列可供科学研究者和管理者参考的结论。

Productions of scientific research should be automatically classified. This thesis mainly used the method of text data mining to realize these functions. Data mining was introduced to automatically categorize master thesis via abstracts and keywords data. More than two thousand data was collected. On the basis of analyzing the characteristics of data, the catego- rization algorithm was determined as Support Vector Machine （ SVM ） , and its details of construction process were ana- lyzed. Finally, simulation of the collection data was made, and the SVM classifier was compared with other common classi- fiers, which shows that SVM has a higher accuracy rate than other commonly used methods. As a conclusion, simulation results about the knowledge was analyzed, which provide useful reference information for scientific researchers and managers.