B-Spline Approximation of Probability Density Function with Complex Distribution

概要

論文の詳細を見る
A probabilistic method that can directly estimate the probability density function (pdf) of a random variable from data expressed by a set of samples or a histogram is presented. In the method, pdf is approximated by the weighted mixture of a series of B-spline functions. The appropriate weights of each spline function and the number of spline functions which compose the pdf are both determined by the Akaike's information criterion (AIC). An iteration technique is developed for the optimization of the weights. The approximation procedure does not need any prior knowledge on the distribution form, and pdf with complex shape can be approximated by the method.
社団法人日本船舶海洋工学会の論文
1994-09-04