[9a-S301-1~7] 23.1 Joint Session N "Informatics"

10:15 AM - 10:30 AM JST(1:15 AM - 1:30 AM UTC)Presentation by Applicant for JSAP Young Scientists Presentation Award

[9a-S301-4] Explainable Machine Learning for Elucidating Structure–Property Relationships in Carbon Nanotube Fibers

〇Daisuke Kimura¹, Naoko Tajima¹, Toshiya Okazaki¹, Shun Muroga¹ (1.AIST)

Keywords:

materials infomatics,carbon nanotube,fibers

In this study, we investigated the structure–property relationships of carbon nanotube (CNT) fibers prepared from aqueous dispersions using machine learning. Factor analysis, regression analysis, and explainable AI techniques were combined to quantify how multi-stage processing and multiscale structural features influence material properties. The analysis showed that local agglomeration plays a dominant role in determining mechanical strength, whereas the effective CNT length and crystallinity strongly contribute to electrical conductivity. This approach enables interpretable, data-driven evaluation of complex structural effects across different scales.