Abstract
Time-of-flight secondary ion mass spectrometry （TOF-SIMS） data are generally so complex that multivariate analysis such as principal component analysis （PCA） and multivariate curve resolution （MCR） are often necessary to interpret TOF-SIMS data. Interpreting more complex TOF-SIMS data requires further data analysis methods using machine learning and deep learning. In this study, the application of autoencoder which is one of the unsupervised methods based on artificial neural networks into TOF-SIMS data of three polymers was evaluated.
