In the case of datasets of important dimensions, the data is previously transformed into a reduced series of representation functions. This process of transforming the input data into a set of functionalities is called features extraction. This is because the extraction of the characteristics proceeds from an initial series of measured data and produces derived values that can keep the information contained in the original dataset, but discharged from the redundant data.
In this way, the subsequent learning and generalization phases will be facilitated and, in some cases, this will lead to better interpretations. It is a process of extracting new features from the original features, thereby reducing the cost of feature measurement, which boosts classifier efficiency. If the features are carefully chosen, it is assumed that the...
