This study aims to reveal multidimensional data structures in the learning process using the topological data analysis method Mapper. Unlike conventional dimensionality reduction methods such as principal component analysis, Mapper visualizes data while preserving data-specific topological features. The results obtained from Mapper depend on the choice of filter functions, parameter settings, and clustering algorithms. Our previous study investigated the optimal parameter settings for educational data analysis. This paper reports the results of analyzing actual data based on those configurations.

Authors: Hitoshi Inoue, Koichi Yasutake, Osamu Yamakawa, Takahiro Tagawa, Takahiro Sumiya

Published in: International Conference on Information Society (i-Society-2024)

• Date of Conference: 26-28 August, 2024
• DOI: 10.20533/iSociety.2024.0004
• ISBN: 978-1-913572-72-3
• Conference Location: Churchill College, Cambridge, UK