Machine Learning Maps Structural Shifts in Mathematics Problem-Solving Research

A bibliometric study applying machine learning has mapped the evolution and intellectual structure of problem-solving research in mathematics education, revealing a dynamically organized field structured around 16 distinct topics. Crucially, the thematic analysis indicates a growing emphasis on social dimensions, characterized by a pronounced shift from the study of individual cognitive processes towards in-service teacher learning, collaborative settings, and the integration of emerging technologies.

Published in SN Social Sciences, the research was conducted by Erdem Çekmez to address the absence of a comprehensive, macro-level intellectual overview of how problem-solving scholarship has developed within mathematics education. To evaluate publication trends, prominent contributors, collaboration networks, and overarching thematic shifts, the author examined a dataset of 637 research articles published between 2000 and 2025 and indexed in the Web of Science database.

The investigation utilized a bibliometric framework augmented by a hybrid structural topic modeling procedure, combining quantitative performance metrics with qualitative thematic interpretation. In evaluating contributions across the 16 identified topics, the analysis highlighted the United States as leading the field in overall research productivity.

In addition to documenting the shift toward collaborative settings and educator development, the study identified affective dimensions and early years problem solving as notable emerging areas for future scholarly investigation. By delineating the intellectual trajectory of the domain, the author outlined persistent structural gaps and imbalances to help guide upcoming academic inquiry in mathematics education.

Subject of Research: Bibliometric analysis of mathematics education problem-solving research

Article Title: Uncovering the intellectual structure of problem solving: a machine learning-assisted bibliometric approach

Article References: Çekmez, E. (2026). Uncovering the intellectual structure of problem solving: a machine learning-assisted bibliometric approach. SN Social Sciences, 6(11), Article 625. https://doi.org/10.1007/s43545-026-01938-z

Image Credits: AI Generated

DOI: 10.1007/s43545-026-01938-z

Keywords: mathematics education, problem solving, bibliometrics, structural topic modeling, machine learning, teacher learning, collaborative learning, educational research, Uncovering, intellectual, structure, problem