Data augmentation

Referring to the process of artificially creating new data from existing ones, data augmentation is primarily utilized in training machine learning models. This statistical technique enables maximum likelihood estimation from incomplete data, thereby enhancing the model's performance by providing a more diverse dataset for training.