作者: Shoufan Fang , George Z. Gertner , Svetlana Shinkareva, , Guangxing Wang , Alan Anderson
关键词:
摘要: The Fourier amplitude sensitivity test (FAST) can be used to calculate the relative variance contribution of model input parameters predictions made with functional models. It is widely in analyses complicated process modeling systems. This study provides an improved transformation procedure for non-uniform distributions that represent parameters. Here it proposed cumulative probability instead density when transforming FAST. improvement will increase accuracy by reducing errors, and makes more convenient practice. In evaluation procedure, was demonstrated have very high comparison currently use.