You will find there's near relationship among machine learning and compression. A system that predicts the posterior probabilities of a sequence given its overall background can be employed for ideal details compression (by utilizing arithmetic coding on the output distribution). Illustration of linear regression on a knowledge established Regression Evaluation https://www.youtube.com/channel/UC5e34pt2PeSG0OpdaAawnqA/Ventura-IT