Statistical model for electrical energy production

To estimate the amount of electrical energy production to reach the desired service level — answered demand divided by the total demand — a statistical model is required. Model selection for a dataset with no statistical model has been done by many researchers, among which, Vuong proposed a test that the two competing models are equally close to the true statistical model against the alternative hypothesis that one of them being superior to the other. In this article, Vuong’s test is expanded into three models based on Kullback–Leibler criterion and likelihood ratio function. The proposed test has been tested on electrical energy production data in Denmark in 2014, and based on the performed test, Gamma model, Log-Normal model and Generalized Rayleigh model are superior to other models. Our data was acquired from NPS.