Study on Document Measurement and Visual Analysis of Big Data Education Evaluation

Liang Yang, Xu Luo, Ao GaoProceedings of the 2022 2nd International Conference on Education, Information Management and Service Science (EIMSS 2022) 2022 · DOI · OpenAlex

Abstract

Machine learning, recommendation algorithms, and other emerging big data technologies have become essential tools for learning analysis and educational evaluation. As educational evaluation transitions toward more scientific, multi-dimensional, and real-time methodologies, understanding the research landscape becomes crucial. This paper utilizes bibliometric methods to analyze 391 articles indexed by the Web of Science from 2012 to 2022. Using CiteSpace software, the study generates visual knowledge maps to clarify the current research status and development context, concluding that big data technology is of significant importance for the future of educational evaluation.

Introduction

Educational evaluation—the systematic process of judging the effectiveness of teaching and learning—has historically been constrained by small-sample data and manual analysis. The integration of big data promises a revolution in objectivity and depth, making the process more scientific and real-time. However, because this field bridges pedagogy and computer science, its research landscape is often fragmented. This study addresses this "information overload" by identifying major global players, foundational literature, and emerging technological "hot spots" using visual analysis tools.

Methodology

The researchers employed bibliometrics to quantitatively analyze academic literature. Two primary tools were used: CiteSpace 6.1.R2 for generating "knowledge maps" (including co-citation and keyword emergence analysis) and Excel 2019 for statistical tracking. The dataset was curated from the Web of Science (WOS) Core Collection, specifically targeting "Articles or Proceedings Papers" at the intersection of "education evaluation" and "big data" published between 2012 and 2022. After removing duplicates and irrelevant entries, 391 valid records were analyzed.

The research volume has seen a dramatic increase over the last decade. Starting with only 6 papers in 2012, the field reached a peak of 60 documents in 2019. While research began a steady climb in 2012, 2017 marked a critical turning point. The authors describe this as an "outbreak period" where the research shifted from theoretical exploration to intense, practical application.

Annual publication volume of documents related to big data in educational evaluation (2012–2022).

Geopolitical and Institutional Landscape

The analysis reveals a distinct geopolitical divide. China is the undisputed leader in terms of publication volume with 160 articles, followed by the USA (45) and Spain (22). However, the United States maintains the highest citation impact (295 total citations), indicating that US-based research currently serves as a more frequent foundational reference for the global community. The study also notes that while China leads in volume, its breadth of international cooperation is slightly lower than that of the United States.

Knowledge Mapping and Research Frontiers

The study identified six primary research clusters that define the current state of the field. These clusters represent the "knowledge base" and research frontiers of big data in education:

  1. Pre-service Teachers: Focusing on NLP and deep learning for teacher training.
  2. Education: General AI and learning analytics applications.
  3. Web Scraping: A cluster driven by the need for data collection during the COVID-19 pandemic.
  4. Sustainability: Focusing on the longevity and ethics of e-learning algorithms.
  5. Information: Educational data mining and knowledge graphs.
  6. Site Selection: Infrastructure and platform selection for academic big data.
Timeline visualization of the primary research clusters, showing the evolution and duration of key topics from 2012 to 2022.

Technological convergence is evident in the rise of deep learning for sentiment analysis (evaluating student emotions) and collaborative filtering for personalized learning recommendations.

Conclusion and Future Directions

The bibliometric analysis confirms that big data is reshaping educational evaluation into a more scientific discipline. The next frontier is expected to be "emotional analysis" via deep learning to better understand the human element within massive datasets. Limitations of this study include the restriction to the Web of Science database and the inherent time lag in publication indexing for the most recent year (2022). Future work should aim for broader international cooperation to bridge the gap between high-volume production and global research impact.

Summary updated

Abstract

Machine learning, recommendation algorithm and other new big data technology, has become an important tool for learning analysis education evaluation.Educational evaluation is an important means to evaluate the value of educational activities, and big data technology is becoming an important part of educational evaluation.It has become an urgent problem to sort out the status quo of big data education evaluation research and clarify the context of its research development.In this paper, we use bibliometrics to analyze the data of 391 articles indexed by Web of Science in recent ten years, and use Cite Space software to generate visual results.Through the analysis of the results, it is concluded that big data technology is of great significance for educational evaluation.
BibTeX
@inproceedings{yang2022study,
  author = {Liang Yang and Xu Luo and Ao Gao},
  title = {{Study on Document Measurement and Visual Analysis of Big Data Education Evaluation}},
  booktitle = {Proceedings of the 2022 2nd International Conference on Education, Information Management and Service Science (EIMSS 2022)},
  year = {2022},
  doi = {10.2991/978-94-6463-024-4\_53},
}
Figure