Abstract: This paper takes the literature related to the computer engineering practice training mode on CNKI as a sample, conducts a quantitative analysis of the literature from the dimensions of high-impact authors, high-frequency keywords, and time-space distribution, sorts out the research context of my country's engineering practice training mode, and reveals this field. The research hotspots and research trends of this paper provide a reference for the further development of related research on the engineering practice training mode of computer courses.
Introduction
The rapid evolution of emerging technologies—including artificial intelligence, quantum computing, and big data—has fundamentally altered the global industrial landscape. This shift has created an urgent demand for computer science graduates who possess not only theoretical knowledge but also robust, hands-on engineering skills. In China, this educational imperative is codified in national strategic goals such as the 14th Five-Year Plan. To effectively refine how students are trained, it is essential to map the current research landscape of "engineering practice training modes." This paper utilizes bibliometric methods to analyze the research context, identify dominant trends, and highlight systemic weaknesses in the existing literature.
Methodology
The study employs a combination of bibliometrics and visual knowledge mapping to analyze academic literature from the China National Knowledge Infrastructure (CNKI) database. The authors queried the database for papers published between 2000 and 2022, specifically targeting the "engineering practice training mode" within computer disciplines. After rigorous data cleaning, a sample of 287 highly relevant documents was isolated. These documents were then analyzed using VOSviewer, a specialized software tool for constructing and visualizing bibliometric networks. The analysis focused on co-occurrence networks and density maps to evaluate relationships between high-impact authors, research institutions, and high-frequency keywords.
Results
Publication Trends
The analysis reveals that academic interest in computer engineering practice training was virtually non-existent before 2005. As illustrated in Figure 1, the field saw a significant surge in publication volume between 2008 and 2013, eventually peaking in 2016. This trajectory reflects a delayed but aggressive academic response to the tech industry's escalating demand for practical engineering expertise.
Keyword Evolution and Research Hotspots
From a total of 711 keywords, the VOSviewer analysis distilled 87 high-frequency terms into 11 distinct clusters. The dominant research hotspots identified include school-enterprise cooperation, CDIO (Conceive, Design, Implement, Operate), software engineering, and "New Engineering."
As shown in Figure 2, the evolution of these keywords tracks closely with national policy shifts. The focus has transitioned from the "Excellent Engineer Training Program" (a 2010–2020 initiative) to the international CDIO framework, and most recently to "New Engineering." This current phase emphasizes multi-disciplinary integration designed to serve the needs of future technology industries.
Discussion
The Collaboration Paradox
A striking finding of the visual mapping is the lack of collaboration among educational researchers. While the literature heavily promotes "collaborative" education between universities and enterprises, the researchers themselves remain siloed. The visual maps reveal highly scattered author networks with almost no long-term, stable cooperative relationships or central thought leaders. This fragmentation suggests a lack of unified direction in the field.
Superficiality and Implementation Gaps
The authors offer a candid critique of the current state of research, noting that it suffers from significant homogeneity. Many studies remain on the surface of the problem, offering repetitive, low-level analyses rather than proposing fundamental, innovative changes to course structures. Furthermore, the literature highlights persistent implementation gaps, including outdated course content, poor docking mechanisms with enterprises, and a lack of robust evaluation metrics.
Conclusion and Future Directions
While this study is limited by its reliance on the CNKI database and a relatively small sample size, it provides a critical overview of the Chinese domestic landscape for computer engineering education. To move beyond the repetitive application of existing models, the authors propose the development of a "three-level implementation system" for industry-university cooperation. Future efforts must focus on creating a robust, long-term ecology for multi-subject collaborative education to meet the demands of the rapidly evolving tech landscape.
Summary updated