基于隐马尔可夫模型的技术融合趋势识别研究
Trend Recognition of Technology Convergence Based on Hidden Markov Model
刘鑫慧 1康乐乐 1魏铭辕1
作者信息
- 1. 南京大学数据智能与交叉创新实验室、信息管理学院,江苏南京 210023
- 折叠
摘要
[目的/意义]识别技术融合状态并把握技术融合的发展趋势有助于更准确地进行创新发展,抓住新兴技术机会.[方法/过程]本研究采用隐马尔可夫模型(Hidden Markov Model,HMM),基于PATSTAT专利数据,依靠技术融合的相似性特征与互补性特征对35个技术领域的技术融合状态进行识别.[结果/结论]分析发现,技术融合状态主要分为封闭、低跨度、高跨度、开放,四种不同的类型.随着时间的推移,不同的技术领域展现出不同状态间切换的特征.从低跨度状态转移到高跨度或开放状态最为常见,体现了技术的多元融合趋势.[创新/局限]本研究首次采用HMM模型从动态视角刻画技术融合状态并追踪其长期趋势,但模型的实际应用需要根据场景进一步构建,后续研究可进一步进行相关探索.
Abstract
[Purpose/significance]Identifying the state of technology convergence(TC)and grasping the development trend of technol-ogy fusion is helpful in conducting innovation activities and benefiting from emerging technology.[Method/process]This study ana-lyzed the PATSTAT patent data to identify the technology convergence in 35 technological fields.Specifically,Hidden Markov Model(HMM)was adopted to estimate the hidden states of technology convergence based on the observations of complementary technology convergence and substitutability technology convergence.[Result/conclusion]The analysis results show that the hidden status of the technology convergence state has four types,namely closed,low-span,high-span and open.Over time,technological fields exhibit the characteristics of switching between different states.The transition from low-span state to high-span state or open state is the most common,reflecting the trend of integrating multiple technologies.[Innovation/limitation]The study innovatively adopts the HMM model to depict the state of technology convergence from a dynamic perspective and track its long-term trend.However,the actual ap-plication of the model needs to be further constructed according to the scene,and relevant exploration can be further conducted in sub-sequent studies.
关键词
互补性技术融合/相似性技术融合/专利/HMM/动态分析Key words
complementary technology convergence/substitutability technology convergence/patent/HMM/dynamic analysis引用本文复制引用
基金项目
国家自然科学基金面上项目(72072087)
国家自然科学基金青年科学基金(71802017)
出版年
2024
NETL
NSTL
万方数据