Abstract
We sometimes use the statistical model for truncated data when we want to predict the final stage for a pandemic or failures using the data in early pandemic period or failures. In such a situation, the estimates for number of patients or failures via the likelihood principle often become smaller than the true ones, which could be the dangerous side. Here, we consider the case of the Weibull distribution for the underlying probability distribution function, and compare the results among the statistical truncated model, the simple ordinary differential equation model, and the SIR model which expresses the pandemic models.