Predicting and following T1 events in dry foams from geometric features

Machine learning techniques have been recently applied in predicting deformation amorphous materials. In this study, we extract structural features around liquid film vertices from images of flowing 2D foam and apply a multilayer perceptron to predict local yielding. We evaluate their importance the description T1 events show that high level predictability may be achieved using well-chosen combinations as prediction data. The most relevant are extracted by performing predictions separately for isolated sets features, these findings verified principal component analysis. Using approach, determine which properties important with regard physics processes. Our indicate lengths angles between films joining at vertex yield events. These two describe 83% As an application, statistics event waiting times experiment.