An Extension of the Levenshtein Distance for Learning Process Analysis in Constructive Learning and an Evaluation of its Effectiveness

In this study, we propose an applied extension of the Levenshtein distance for analyzing learning logs in constructive programming learning. By incorporating the influence of blank and dummy cards—which conventional distance metrics fail to adequately capture—our method enables a more detailed evaluation of the quality of learners’ trial-and-error processes and the presence or absence of logical thinking. This approach clarifies the distinction between exhaustive trial behaviors and logically constructed problem-solving, thereby contributing to learning support systems that can provide more personalized and appropriate feedback.