About this Abstract |
Meeting |
2023 TMS Annual Meeting & Exhibition
|
Symposium
|
AI/Data Informatics: Computational Model Development, Validation, and Uncertainty Quantification
|
Presentation Title |
Prediction of Glass Transition Temperature by Machine Learning Method with Soft Constraint |
Author(s) |
Jin Myoung Jeon, Tae-Min Yeo, Jung Wook Cho |
On-Site Speaker (Planned) |
Jin Myoung Jeon |
Abstract Scope |
In this study, soft constrained neural network model was developed to predict glass transition temperature with considering mixed modifier effect. Iterative Gaussian process regression was used to remove outlier in glass system database. Principal component analysis was used to analyze the effect of experimental conditions on database. By controlling the weight of penalty term, it was able to obtain the model which shows good performance with low inequality loss. The potential of soft-constrained model was confirmed by comparing with no-constrained model for mixed modifier composition cases. |
Proceedings Inclusion? |
Planned: |
Keywords |
Other, |