Kernel

A kernel is a defined neighborhood (area) of interest around a point or area object. Mathematically, as in the classic example of kernel density estimation (KDE), it is defined by a kernel function that has two characteristics. First, each kernel function specifies a range over which it is to be evaluated, using either simple connection, distance, or a count of cells (pixels) in a raster. In KDE, this range is referred to as the bandwidth of the function. Choice of bandwidth is similar to the choice of bin width when compiling a histogram, with large values resulting in smooth density estimate surfaces, and vice versa. Second, it provides some weighting that specifies the effect of distance on the calculation to be performed.