I am fitting a spline to some data using mgcv in R. I "know" that most of the variation should be in the first half of the range of the independent variable. The other half should very smooth, but is sometimes overfit due to "noise".
Is there any way to control the smoothness in specific parts of the spline? E.g. by penalizing wiggliness harder in part of the spline?
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…