一种基于机器学习的专利可转让性评估方法研究被引量：5

A Patent Transferability Evaluation Method Based on Machine Learning

下载PDF

导出

摘要 [研究目的]面对海量的专利数据,如何构建有效的专利可转让性评估方法,筛选出具有转让可能性的专利,对于发现潜在高价值专利、提升我国专利成果转移转化效率具有重要意义。为此,提出一种基于机器学习的专利可转让性评估方法。[研究方法]首先利用基于机器学习的指标约减算法对从技术维度、法律维度、经济维度和主体维度构建的专利可转让性评估指标进行约减,以去除掉冗余指标;然后根据约减后得到的评价指标构建机器学习模型,并对专利可转让性进行评估。[研究结论]人工智能领域专利可转让性评估案例结果表明,当剔除掉冗余指标后机器学习模型的预测准确率均值提升了0.56%;基于机器学习的专利可转让性评估模型的分类准确率达到72.36%,可以较好地对专利的可转让性进行评估。案例结果验证了所提出方法的可行性和有效性,该方法为专利可转让性评估提供了新的研究方法。 [Research purpose]In the face of massive patent data,it is of great significance to construct an effective patent transferability assessment method and screen out patents with possibility of transfer,which is of great significance for discovering potential high-value patents and improving the transfer and transformation efficiency of patents in China.Therefore,a method of patent transferability evaluation based on machine learning is proposed.[Research method]Firstly,the indicator reduction algorithm based on machine learning is used to reduce the patent transferability evaluation indicator constructed from the technical dimension,legal dimension,economic dimension and subject dimension to remove redundant indicators.Then,a machine learning model is constructed according to the evaluation indicators obtained after the reduction,and the transferability of the patent is evaluated.[Research conclusion]The results of the patent transferability evaluation case in the field of artificial intelligence show that the average prediction accuracy of the machine learning model increases by 0.56%when redundant indicators are removed;the classification accuracy of the patent transferability evaluation model based on machine learning reaches 72.36%,which can better evaluate the transferability of patents.The case results validate the feasibility and effectiveness of the proposed method,which provides a new research method for patent transferability assessment.

作者李欣冯野马晓迪 Li Xin;Feng Ye;Ma Xiaodi(College of Economics and Management,Beijing University of Technology,Beijing 100124)

机构地区北京工业大学经济与管理学院

出处《情报杂志》 CSSCI 北大核心 2023年第3期85-93,共9页 Journal of Intelligence

基金国家自然科学基金面上项目“基于多源数据融合与深度学习的颠覆性技术早期识别方法研究”(编号:72174017)。

关键词专利可转让性机器学习专利可转让性评价指标指标约减算法人工智能 patent transferability machine learning patent transferability evaluation indicators indicator reduction algorithm artificial intelligence

分类号 G306 [文化科学]