Definition: Unnormally is a term used in machine learning and statistics. It refers to a technique or procedure where data points are normalized so that they have zero mean and unit variance, which is often done for training purposes before using them as features in a machine learning model. This normalization helps in reducing the bias introduced by large differences between observations when considering them as features. Unnormally normalized data can be useful for various applications such as feature selection, clustering, and dimensionality reduction.