Dimension Identification, Feature Extraction and Layered Model of Digital Economy Security: A Mixed Research of LDA Thematic Analysis and Grounded Theory Coding
Abstract：Digital economy security is an important support for modernizing national security systems and capabilities. The existing studies have carried out theoretical exploration from the perspective of multiple disciplines but lack dimensional identification as well as structural feature analysis of digital economy security. This paper adopts a mixed research method to address this problem. Study 1 is a text analysis of digital economy security. Based on 1,838 articles from 62 top international journals, a Latent Dirichlet Allocation (LDA) topic model is constructed in Study 1. The results show that the Western academic community divides the digital economy security into five categories: digital infrastructure risk, key core technology risk, digital industry security, industrial digital security, and digital social stability, showing different topic features. Social network analysis further shows that these five types of security present a core-peripheral layered structure, among which the digital infrastructure risk is located in the core layer, digital social stability is in the peripheral layer, and the other risks are in the middle layer.
Study 2 is an empirical research of digital economy security in China. Through semi-structured in-depth interviews with 42 practitioners, we construct 5 types of main risks and 18 types of sub-risks based on grounded theory coding. The Delphi method of risk ranking further shows that, China mainly faces five security categories according to the importance of security: key core digital technology security, digital security, digital industry security, industrial digital security, and digital economy ecological environment. Among them, key core digital technology security is represented by digital technology design security, digital technology manufacturing security, digital technology supply security and other features. Digital security includes data security, cyber security, information security, artificial intelligence security and other features. Digital industry security is characterized by digital industrial chain and supply chain security, platform enterprise monopoly, digital product price discrimination, digital industry competitiveness and so on. Industrial digital security is manifested in the features of digital transformation of manufacturing industry, digital agriculture construction, and digital financial security. The features of digital economy ecological environment include digital infrastructure environment, digital market environment, digital policy environment, digital social environment. Study 2 further verifies and modifies the conclusions of Study 1.
According to the comparative results of Study 1 and 2, the manifestation of digital economy security in China is not quite different from that in the West. However, in terms of risk ranking, China pays attention to the key core digital technologies security, while the West focuses more on digital infrastructure security. The West regards digital social stability as the peripheral layer of digital economy security, while China integrates the ecological environment of digital economy into the security system. It should be noted that security types with Chinese characteristics, such as artificial intelligence security, digital industrial chain and supply chain security, digital industry competitiveness, digital agriculture construction and digital infrastructure environment, have also been identified.
This paper has three major theoretical contributions. First, it deeply reveals the dimension types and topic features of digital economy security and builds a layered conceptual model. Although the topic of digital economy security has attracted more attention in the recent years, its internal dimension form and relationship structure are still a black box, with a lack of detailed and in-depth theoretical elaboration. This research provides a more fine-grained explanation by dividing the five dimensions of digital economy security, revealing the preference ranking of risk, indicating that it is no longer regarded as a homogeneous and unitary whole, and expanding the research on the typology structure of digital economy security. Second, comparing the differences between China and the West in digital economy security it clarifies the direction for future researches in China. At present, the international academic community has carried out a large number of exploratory research but the domestic research is still in the initial stage, so it is necessary to keep up with the theoretical frontier and grasp the academic innovation. Our results reveal the dimensional division of 62 international top journals, which provides implications for future researches. At the same time, combined with the Chinese situation and local practice, the study refines the security elements of digital economy with Chinese characteristics, and promotes the dialogues between international theoretical literature and Chinese practice. Third, the mixed research methods are adopted to make up for the limitations of the existing literature focused on speculative discussion, and for the first time to provide empirical evidence as well as support for the conceptual model of digital economy security. Study 1 uses the LDA topic model of unsupervised machine learning to overcome the subjective bias brought about by traditional manual coding by automatically extracting potential topics from massive data and provide evidences for dimension identification. Based on the Chinese reality, Study 2 abstracts new security dimensions through grounded theory. These two studies confirm each other, improving the external validity of the conclusion, promoting the dialogues between Chinese and Western theories. At the same time, they complement each other, bridging the gap in theory and practice, and contributing the knowledge increment to the construction of digital economy security theory with Chinese characteristics.
From the practical level, the layered conceptual model constructed in this research is conducive to helping government officials and enterprise managers to better conduct hierarchical governance of digital economy security. For the core digital technology security in the core layer, we should give full play to the advantages of the new national system, concentrate on building a large-depth, interdisciplinary collaborative research mechanism to solve the bottleneck problem. For the security in the peripheral layer, we should keenly capture weak signals in digital ecological environment and provide theoretical guidance for government agencies to regularly issue dynamic warning of digital economy risks and establish a scientific and effective security monitoring system.
范柏乃 盛中华. 数字经济安全的维度识别、特征提取及分层模型——基于LDA主题分析与扎根理论编码的混合研究[J]. 浙江大学学报(人文社会科学版), 0, (): 1-.
Fan Bonai Sheng Zhonghua. Dimension Identification, Feature Extraction and Layered Model of Digital Economy Security: A Mixed Research of LDA Thematic Analysis and Grounded Theory Coding. JOURNAL OF ZHEJIANG UNIVERSITY, 0, (): 1-.
