基于多维标度法的专利文本可视化聚类研究

Research on MDS-based visual clustering in patent texts

为了从当前海量的专利文本信息资源中迅速准确地获取所需的信息并将其以可视化的形式展现出来,通过研究文本挖掘相关理论和关键技术,结合专利文本的特征及现有的分析方法,提出一种基于多维标度法的专利文本可视化聚类方法,并结合实际进行了相关分析讨论。实验结果表明,该方法较好地展现了当前专利技术领域的应用主题分布状态及其关联图谱,可为相关部门的管理决策和技术创新提供有意义的参考和支持。

In order to acquire the information need from the current large amounts of patent texts and make them visualized to people quickly and exactly,this paper proposed a kind of visualized clustering method in patent texts based on MDS through researching the correlative theory and key technology of text mining,combined with the character of patent texts and existing analysis methods. It also made some relative analysis and discussion about the application of this method. The experiments show that this method displays the current distribution of patent texts application topics well; it can also effectively support the decision in management and technology innovation.