Research on the Construction of Evaluation Indicators of Students' Computational Thinking Based on Spectral Clustering

The scientific and reasonable evaluation indicators that can fully reflect the various dimensions of students' computational thinking (CT) skills is the basis and premise of accurately evaluating students' CT skills, which is of great significance to the cultivation of students' CT skills. However, the current research on how to construct the evaluation indicators is still inadequate, and most of the research is put forward by the subjective experience of researchers, lacking objectivity and the universality of ability. In the paper, we comprehensively reviewed the concepts of CT in the theoretical literature of CT, aiming to construct the comprehensive and effective evaluation indicators of CT for students by clustering the keywords of CT concepts and extracting indicators. The validity of indicators is verified by qualitative analysis, quantitative analysis and expert evaluation. The results show that the evaluation indicators of CT constructed by spectral clustering technology are a more scientific, more comprehensive reflection of the ability dimensions of CT. It has unique advantages in constructing objective and comprehensive evaluation indicators and provides an evaluation basis for the evaluation practice of CT skills.