Generalization Ability

It refers to the performance (ability) to successfully predict not only the given training data but also unknown new data during model training.

For example, if a model is created to predict the weather, it is useless if the model can predict the weather in the past, but not tomorrow's weather.
Such a situation is called low generalization performance.
For this reason, it is very important for machine learning models to improve their generalization performance (ability), and various methods are being researched for this purpose.



Made with
by VARISTA Team.
© COLLESTA, Inc. 2021. All rights reserved.