Modelização fracionária do comportamento de atenuação linear de metais puros

Abstract

The linear attenuation coefficient values obtained numerically were adjusted to a fractional Grünwald-Letnikov model implemented in Python, in order to evaluate its ability to describe the behavior of radiation in different metals. The fit is analysed through the parameters R², RMSE, χ², AIC and BIC, complemented by statistical tests for the diagnosis of residuals such as the Shapiro–Wilk test (normality), Breusch–Pagan (homoscedasticity) and Durbin–Watson (autocorrelation). At the same time, Ridge regression was applied as a regularization method for the eight metals studied.