The use of machine learning and AI is being promoted in the research and development of lubricant additives. In addition to conventional experimental verification using friction tests and surface chemical analysis using lubricant additives, analytical verification using machine learning, such as modeling of friction phenomena, makes it possible to investigate the mechanism of tribofilm formation from lubricant additives and their friction and wear properties from a multifaceted and comprehensive perspective. In this study, we aim to elucidate the lubrication mechanism by extracting the importance of feature values in tribofilm formation and its friction and wear characteristics from various lubricant additives using AI such as machine learning and deep learning.