基于混沌理论和专利挖掘的技术不连续演化特征研究  

Research on the Characteristics of Discontinuous Technological Evolution Based on Chaos Theory and Patent Mining

作　　者：任海英[1] 梁宇航 Ren Haiying;Liang Yuhang(Beijing University of Technology,Beijing 100124)

机构地区：[1]北京工业大学经济与管理学院,北京100124

出　　处：《情报杂志》2025年第4期181-189,共9页Journal of Intelligence

基　　金：北京市社会科学基金重点项目“科技创意及其智能化管理系统的基础理论与方法研究”(编号:23GLA009)研究成果。

摘　　要：[研究目的]针对技术不连续演化定量研究的不足,结合混沌理论与专利挖掘方法,揭示技术不连续演化的混沌特征,在技术不连续演化阶段为技术预测提供混沌分析视角。[研究方法]以语音识别技术为例,首先检索专利数据并构建申请数据时间序列,通过关联维数和李亚普洛夫指数识别技术演化系统的混沌特征及其局部混沌期。之后,对不同时间窗口的数据构建LDA主题模型和共词网络,分析技术主题及相关节点在局部混沌期前、中、后的演化特征,并预测潜在新兴技术。[研究结果/结论]实验结果表明,语音识别技术演化过程中,技术不连续演化阶段与局部混沌期高度重合,为“混沌是技术不连续演化的内在特征”提供实证依据。此外,新兴技术在局部混沌期内产生,并表现为技术主题的稳定发展和相关节点中心性的提升。不仅反映新兴技术对语音识别技术系统的深远影响,而且揭示其未来演化方向。[Research purpose]Aiming at the deficiency of quantitative research on discontinuous technological evolution,this paper combines chaos theory and patent mining methods to reveal the chaotic characteristics of discontinuous technological evolution and provide a chaotic perspective for technology prediction in the period of discontinuous technology evolution.[Research method]Taking speech recognition technology as an example,we first retrieve patent data and construct time series of application data,and identify the chaotic features and local chaotic periods of the technological evolution system through correlation dimension and Lyapunov exponents.Then,LDA topic models and co-word networks are built on data from different time windows,and the evolution characteristics of technical topics and related nodes before,during and after the local chaotic period are analyzed,and potential key technologies are predicted.[Research result/conclusion]In the evolution process of speech recognition technology,the discontinuous technological evolution periods are highly overlapped with the local chaotic periods,providing empirical support for the claim"chaos is the intrinsic feature of discontinuous technological evolution".In addition,key emerging technologies appear during the local chaos period,and are represented by the steady development of technical topics and the improvement of the centrality of related nodes,reflecting the far-reaching impact of key emerging technologies on speech recognition technology systems and revealing their future evolution direction.

关 键 词：混沌理论 专利文本 专利挖掘 技术演化 主题模型 共词网络 语音识别技术 关联维数 李亚普洛夫指数 

分 类 号：G35[文化科学—情报学]