Another motivation for feature selection is that, since our goal is to approximate the underlying function between the input and the output, it is reasonable and important to ignore those input features with little effect on the output, so as to keep the size of the approximator model small.For example, [Akaike, 73] proposed several versions of model selection criteria, which basically are the trade-offs between high accuracy and small model size.
đang được dịch, vui lòng đợi..
