 
 Red crosses show the locations of 10 randomlyselected observations from a lognormal distribution.
 The dotted curve is the probability density of that standard normal population, f[y].
 The green curve shows the result of Gaussian smoothing using the 'plug in' estimate of h.
 The Gaussian smoothing function for n observations is the mean normal probability density (or cumulative probability) for n observations of variable Y, at location y, given a standard deviation h.
 Blue and orange curves on the second graph show the results of undersmoothing and oversmoothing.
