Hi all:

A given continuous variable which is not normally distributed, how can I work out the Probability Density Distribution Function? I have been search on this topic for couple weeks. Closest tutorial is on wikipedia about kernel estimation by bayes theorem. That is a diabete study.

But I am not sure that if I use the kernel technique the area under the probability curve can be 1. Sum of the probability density is 1, tell me if I got this wrong.

So far I treat the density function discrete for simulation I working on to get around the non-deterministic function problem. And I suspect that there is actually no very good techniques to estimate this kind of function; since, all baysian network I saw treat probability density function discrete.