Selection of relevant texture descriptors for recognition of HEp-2 cell staining patterns

What's The Big Idea? Ridge Regression and Regularisation

“…When overfitting happens, a model will describe the data it was trained on quite well but is likely to perform poorly on new data. Overfitting can be resolved by making a “ridge‐type” regularisation adjustment, which – as before – shrinks the magnitude of the estimated regression coefficients, leading to a simpler model but with better predictive ability 5–7 . Other examples where the number of variables can be much larger than the number of observations can be found in genetics.…”