专利信息的技术功效与应用图挖掘研究

The Mining Research of Technical Efficiency and Application Map of Patent Information

目前,专利技术功效图与技术应用图的构造方法主要通过专家或学者的主观判断来对技术、功效或应用进行分类,因此比较耗费人力,结果也不够全面。针对以上问题,提出一种基于文本挖掘的、相对客观的解决方案,首先对专利信息进行文本挖掘并从中提取专利文献中涉及的技术、应用和功效,再由专家对文本挖掘后的结果进行评估以确定较全面、准确的技术、功效、应用的特征,得到的特征将能直接用于统计其对应包含文献数量并最终构造技术功效图与技术应用图。实验结果表明,得到的技术功效等特征比专家预想的特征更全面、准确,技术功效图与技术应用图更加完整。

Currently, the research of patent technology effect matrix and technology application map is primarily depend on the judgement of experts in classification of technology, effect and application terms, which requires many labors and the result seems to be not comprehensive. Therefore, the paper proposes a relatively objective solution based on text min- ing. This solution mainly focuses on the extraction of patent information about technology, effect and application terms, and then the experts make decisions based on the results of extraction to define the relatively comprehensive and accurate terms. With these terms, the maps by calculating the numbers of terms and the times of occurrences among them can be built. Experimental results show that the finally statistics are more comprehensive relative to experts＇ ways.