February 2019
Intermediate to advanced
386 pages
9h 54m
English
This kernel has been proposed to minimize the mean squared error, and it has also the property to be very regular (indeed, it can be imagined as inverted parabola). The formula is as follows:

The constant ε has been introduced to normalize the kernel and meet all requirements (in an analogous way, it's possible to extend the kernel in range (-h, h) so as to be more consistent with the other functions). A graphical representation is shown in the following screenshot:

The kernel can become very peaked when ...
Read now
Unlock full access