“…Moreover, some, but not all, studies use measures that contain the word predictive in their name even though they do not employ standard methods from statistics, which fit the model to a subset of the data and use the remaining data to assess the predictive capability of the model. Examples of past studies utilize mean square error , bias , variance , R‐squared , root mean square prediction error , predictive relative error , Kolmogorov–Smirnov test , the Theil statistic , sum of squared error and mean error of prediction , while Inoue et al. considered the Akaike information criterion (AIC) to compare the Jelinski–Moranda model with and without CP.…”