This paper develops an optimized prediction method based on machine learning for optimal process parameters vacuum carburizing. The critical point is data expansion through a few and data, which leads to optimizing carburization in heat treatment. extends the volume by constructing neural network with augmentation presence of small samples. In this paper, database 213 expanded 2116800 predictio...