Suppose there is a distribution that has values sampled on the interval 1-25 with corresponding sample values that have to be smoothed. For example:
n f[n] 3 123 5 234 8 246 10 567 14 253 18 251 24 132
The first column has sample points between 1-25 but does not have samples for all points on the interval. Should interpolation of other points in the distribution be done so that we have approximate values for all points between 1-25 before applying Gaussian smoothing?