Deep Learning-Based Prediction of Physical Stability considering Class Imbalance for Amorphous Solid Dispersions

This research is aimed at predicting the physical stability for amorphous solid dispersion by utilizing deep learning methods. We propose a prediction model that effectively learns from small dataset imbalanced in terms of class. In order to overcome imbalance problem, our performs hybrid sampling which combines synthetic minority oversampling technique (SMOTE) algorithm with edited nearest neighbor (ENN) and reduces dimensionality using principal component analysis (PCA) during data preprocessing. After preprocessing, it process carefully designed neural network simple but effective structure. Experimental results show proposed has faster training convergence speed better test performance compared existing DNN model. Furthermore, significantly computational complexity both processes.