针对原始德温特专利信息内容表示粒度大,不能够满足对该数据直接进行专利统计分析的问题,本文提出了一种基于文本预处理的德温特专利信息分类方法。该方法结合文本预处理技术,通过对专利数据结构和内容的分析,实现了对专利内容进一步分类。实验结果表明,该方法有效的降低专利信息内容的表示粒度,从而提高了专利分析结果的准确性和多样性。 According to the fact that the original Derwent patent information content shows a large granularity and can not satisfy the patent statistical analysis directly, this paper proposes a Derwent patent information classification method based on text preprocessing. The method combined with text preprocessing technology, through the patent data structure and content analysis, to achieve further classification of patent content. The experimental results show that this method effectively reduces the representation granularity of patent information content and improves the accuracy and diversity of patent analysis results.