基于机器学习的专利质量评价研究

随着专利数量的迅猛增长，专利质量问题日益凸显，专利质量评价成为学者关注焦点。面对海量专利数据，如何构建有效的专利质量评价方法，实现对专利质量的准确评价和分类是一项急迫的工作。首先，通过对国内外相关专利质量评价研究进行系统调研和梳理，提出一种新的专利质量评价指标体系；其次，根据新的专利质量评价指标体系，构建一种基于机器学习的专利质量评价方法；最后，以人工智能技术专利为例进行实证研究。结果表明，有效综合利用专利技术性、经济性、法定性和主体性评价指标，有助于更加全面、深入地评价专利质量；基于机器学习的专利质量评价方法能够迅速对专利进行分类并识别高质量专利，从而提高专利质量评价结果深度，为实现大规模专利质量分类评价智能化提供可能。

Research on Patent Quality Evaluation Using Machine Learning

With the rapidly growth of patents, patent quality issues have become increasingly prominent, and patent quality evaluation has become the focus of scholars. Faced with a large number of patent data, how to construct an effective patent quality evaluation method to achieve accurate evaluation and classification of patent quality is an urgent task. Firstly, this paper proposes a new patent quality evaluation indicator system through systematic investigation and review patent quality evaluation research literature. Secondly, according to the new patent quality evaluation indicator system, a patent quality evaluation method based on machine learning is constructed, and the artificial intelligence technology patent is taken as an example for empirical research. The research results show that the effective comprehensive utilization of technical, economic, legal and subjective evaluation indicators of patents is helpful to evaluate the quality of patent in more comprehensive and in-depth; The patent quality evaluation method based on machine learning can quickly classify and identify high-quality patents, thereby improving the depth and value of patent quality evaluation results, and providing the possibility to realize the intelligentization of large-scale patent quality classification evaluation.