Comparison of Machine Learning Approaches for Classifying Spinodal Events
This work provides a performance benchmark for spinodal event classification, which is incremental as it compares existing models on a specific dataset.
The paper compared deep learning models for classifying spinodal events, finding that NAT and MobileViT outperformed others with accuracy up to 94.65% and AUC of 0.98, surpassing a CNN model at 88.44% accuracy.
In this work, we compare the performance of deep learning models for classifying the spinodal dataset. We evaluate state-of-the-art models (MobileViT, NAT, EfficientNet, CNN), alongside several ensemble models (majority voting, AdaBoost). Additionally, we explore the dataset in a transformed color space. Our findings show that NAT and MobileViT outperform other models, achieving the highest metrics-accuracy, AUC, and F1 score on both training and testing data (NAT: 94.65, 0.98, 0.94; MobileViT: 94.20, 0.98, 0.94), surpassing the earlier CNN model (88.44, 0.95, 0.88). We also discuss failure cases for the top performing models.