Distribution Fitting 1

Let be Y = (Y1, ..., Yn) and X = (X1, ..., Xn) series of pair observations and the objective be finding of a function f(x; a1, ..., am) for which Ŷ = f(X) is the best possible solution of the approximation Ŷ ~ Y. Reaching this objective suppose finding of the expression of the f function and of the values of a1, ..., am parameters. Under assumption of agreement between observation and model the expression of the function f is supposed to be known (or at least supposed, when a search from a given set of alternative expressions is conducted). Thus, it is remaining to obtain the values of a1, ..., am parameters. In order to have a unique solution for the values of the a1, ..., am parameters at least m ≤ n is required to be assured. A series of alternatives are available for Y ~ f(X) approximation, that ones considered most important being revised and exemplified in this research.