Research on potential disruptive technology identification based on technology network

Three evident and meaningful characteristics of disruptive technology are the zeroing effect that causes sustaining technology useless for its remarkable and unprecedented progress, reshaping the landscape of technology and economy, and leading the future mainstream of technology system, all of which have profound impacts and positive influences. The identification of disruptive technology is a universally difficult task. Therefore, the paper aims to enhance the technical relevance of potential disruptive technology identification results and improve the granularity and effectiveness of potential disruptive technology identification topics. According to the life cycle theory, dividing the time stage, then constructing and analyzing the dynamic of technology networks to identify potential disruptive technology. Thereby, using the LDA topic model further to clarify the topic content of potential disruptive technologies. This paper takes the large civil UAVs as an example to prove the feasibility and effectiveness of the model. The results show that the potential disruptive technology in this field is the main equipment, data acquisition, and information transmission.

Methods
Through the analysis of the technology life cycle, the division of the patents, the construction of the technology network, the identification of nodes leaping, the clustering of technical topics, we aim to identify potential disruptive technology.

Procedures:

Knowledge flow: being familiar with the technical background knowledge in the field of large civil UAVs, and accomplishing the technical decomposition.
Invention patents: analyzing the technology life cycle by the loget lab to separate the invention patents into four parts. According to each part, constructing the IPC technical network and identifying the leapfrogging and diffusible nodes.
Technical topics: making use of the LDA model to cluster and explain the broad and various content of the inventions.

Testing:
Dividing the inventions of the embryonic stage into two groups and examining them by means of the Mann-Whitney test. Finally, the result shows the huge differences in the patent value, sustaining influence, and creativity for the superiority of disruptive technology.

Software：
loget lab == 4.0
gephi == 0.10.1
python == 3.11.4
spss == 26

颠覆性技术（disruptive technology）具备三项显著且兼具价值的核心特征：其一为归零效应，即颠覆性技术凭借显著且前所未有的技术进步，令持续性技术（sustaining technology）完全失效；其二为重塑技术与经济的整体格局；其三为引领技术体系的未来主流发展方向，上述特征均具备深远影响与积极价值。而颠覆性技术的识别始终是一项全球性的难题。为此，本研究旨在提升潜在颠覆性技术识别结果的技术相关性，并细化潜在颠覆性技术识别主题的粒度与有效性。基于技术生命周期理论划分时间阶段，随后构建并分析技术网络的动态演化特征，以此识别潜在颠覆性技术；进而借助LDA主题模型（LDA topic model）进一步厘清潜在颠覆性技术的主题内涵。本文以大型民用无人机（large civil UAVs）为案例，验证所提方法的可行性与有效性。研究结果表明，该领域的潜在颠覆性技术主要集中于核心装备、数据采集与信息传输三大方向。

研究方法
通过技术生命周期分析、专利阶段划分、技术网络构建、节点跃迁识别以及技术主题聚类，实现潜在颠覆性技术的识别。

流程：
1. 知识梳理：熟悉大型民用无人机领域的技术背景知识，并完成技术分解工作。
2. 发明专利分析：依托loget lab开展技术生命周期分析，将发明专利划分为四个时间阶段；针对每个阶段构建IPC技术网络，并识别跃迁型与扩散型节点。
3. 技术主题分析：利用LDA主题模型对发明专利的多元宽泛内容进行聚类与内涵阐释。

验证环节：
将萌芽阶段的发明专利划分为两组，采用曼-惠特尼检验（Mann-Whitney test）开展显著性验证。最终结果显示，相较于非颠覆性技术，颠覆性技术在专利价值、持续性影响力与创新性层面均存在显著差异。

软件环境：
loget lab == 4.0
gephi == 0.10.1
python == 3.11.4
spss == 26