基于机器学习的成本法在专利价值评估中的应用研究--以“新能源汽车”为例  

Research on the Application of Machine Learning-based Cost Method in Patent Value Assessment--Taking“New Energy Vehicle”as the Case Ran Congjing

作　　者：冉从敬[1] 李旺[1] 胡启彪 黄文俊 Li Wang;Hu Qibiao;Huang Wenjun(School of Information Management,Wuhan University,Wuhan 430072,China)

机构地区：[1]武汉大学信息管理学院,湖北武汉430072

出　　处：《现代情报》2024年第5期140-152,共13页Journal of Modern Information

基　　金：国家社会科学基金重大项目“大数据主权安全保障体系建设研究”(项目编号:21&ZD169);国家社会科学基金青年项目“基于知识元的高校专利质量智能判别及其推荐研究”(项目编号:23CTQ028);国家自然科学面上项目“基于图卷积神经网络的新兴技术领域高质量专利识别及其演化研究”(项目编号:72274084)。

摘　　要：[目的/意义]构建基于机器学习的成本法专利价值评估方法,快速识别海量专利的实际成本,并预测其价值区间,在为专利价值评估提供新研究思路的同时,也为专利转移转化定价提供了参考借鉴。[方法/过程]通过Innography数据库与Incopat数据库下载“新能源汽车”领域多指标专利数据,提取专利成本影响因素与专利价值影响因素,并形成专利数据训练集与专利数据预测集;构建AutoGluon机器学习分类算法,将包含成本数据的Innography专利数据训练集导入模型进行训练,并将训练好的模型对Incopat专利数据预测集进行成本预测;最后使用成本法并结合本研究提出的专利价值指数对预测结果进行计算,估算其价格区间。[结果/结论]通过实证分析与结果验证可知,本研究构建的基于机器学习的成本法专利价值评估方法在预测专利价值区间中具备一定有效性,为促进专利价值评估研究深化及专利转移转化定价实践发展提供了参考。[Purpose/Significance]Constructing a machine learning-based patent value assessment method is constructed to quickly identify the actual costs of a large number of patents and predict their value ranges,which provides a new research idea for patent value assessment as well as a reference for the pricing of patent transfer and transformation.[Method/Process]Using data from Innography and Incopat databases,multiple indicator patent data in the field of“new energy vehicles”were downloaded.The study extracted factors influencing patent costs and patent value.Subsequently,patent data training sets and prediction sets were formed.An AutoGluon machine learning classification algorithm was established,and the Innography patent data training set containing cost data was imported into the model for training.The trained model was then used to predict costs for the Incopat patent data prediction set.Finally,employing the cost approach and combining it with the patent value index proposed in this study,the results were calculated to estimate the price range.[Results/Conclusion]Through empirical analysis and result verification,it was evident that the machine learning-based cost approach for patent valuation constructed in this study demonstrates a certain level of effectiveness in predicting the value range of patents.This provided a reference for promoting the deepening of patent value assessment research and the development of pricing practice in patent transfer and transformation.

关 键 词：机器学习 成本法 价格预估 专利价值 

分 类 号：G250.2[文化科学—图书馆学]