需求驱动的跨领域专利技术挖掘方法构建

The Construction of Demand-Driven Cross-Disciplinary Patent Technology Mining Method

如何从大量的专利信息中快速准确地发掘与企业产品相关的专利技术,仍然是产业界和学术界面临的难题之一。兴奋型需求往往是企业发掘的对象和产品进化的方向,因而,从企业产品创新的需求出发,提出一种需求驱动的跨领域专利技术挖掘方法:通过需求功能映射和功能泛化建立专利检索式,从专利数据库中检索专利,在专利检索式的基础上依据国际专利分类(IPC)进行拓展以保证专利挖掘数据的广泛性,并且通过对专利技术特征、引用关系、法律状态等多个方面的分析,实现对跨领域专利技术的挖掘。该方法使用自然语言处理技术,通过对专利数据的提取和分析,完成跨领域专利技术挖掘、形成了早期方案,并基于技术系统进化理论对所挖掘专利进行创新性评价。通过油烟机实例研究表明,该专利挖掘方法通过功能拓展和IPC分类号的应用,能够全面检索到与目标产品功能相关的专利信息,提高挖掘结果的准确性;且在专利挖掘过程中,引入法律信息、专利引用情况和专利技术文本相似度等指标,并可利用计算机辅助完成大部分工作,能够节省人力资源,提高专利挖掘效率和速度。

In the current paradigm of open innovation,swiftly and accurately identifying external knowledge resources and exploring innovation opportunities become imperative.Patents,as a crucial source of intelligence,represent one of the most significant outcomes of technological research and the most effective carrier of technical information.The essence of patent mining lies in uncovering potential implicit information within patents through various processing techniques such as manipulation,combination,and statistical analysis,ultimately transforming it into intelligence and knowledge serving innovation endeavors.In the realm of product design,demands reflect users'requirements and expectations,constituting the starting and ending points of product design.Excitement-based demands often serve as the objects of exploration for enterprises and the direction for product evolution.Many product innovations,especially breakthrough innovations,fundamentally involve the transfer of cross-disciplinary technological characteristics.However,efficiently and accurately uncovering patent technologies relevant to enterprise products from a vast pool of patent information remains a pertinent challenge.This paper establishes a patent retrieval model through demand-function mapping and function generalization,with an extension based on IPC classification to ensure the comprehensive extraction of patent mining data.It further screens and evaluates patents through multiple dimensions,employs natural language processing technology to assist in extracting technical information,uncovers opportunities for cross-disciplinary technological feature transfer and early-stage solutions,and finally evaluates the early-stage solutions based on the laws of technological system evolution.From this,the paper proposes a demand-driven approach for cross-disciplinary patent technology mining.It utilizes natural language processing technology to extract and analyze patent data,facilitating cross-disciplinary patent technology mining to generate early-stage solutions and creatively evaluate them based on the theory of technological system evolution.Moreover,a case study on cooker hoods demonstrates the high accuracy and efficiency of this method in cross-disciplinary patent technology mining.Besides,most of the patent mining work is completed by computer,which is proved to save manpower,as well as improve the efficiency.