sample size with censored data

The corrected AIC can be computed when the likelihood calculation involves censored values, as long as the effective degrees of freedom are computed in a way that properly accounts for the uncertainty introduced by the censoring.

For example, see equation (14) in https://www.degruyter.com/document/doi/10.2202/1557-4679.1032/pdf which clearly shows this is possible, albeit for a particular example.

Start by reviewing the corrected AIC calculation and the likelihood handling for censored values, then compare the required effective degrees of freedom with equation (14) in the linked paper. Done means corrected AIC can be computed for censored-data likelihoods while accounting for censoring uncertainty; the issue does not name implementation files or tests.

Written by the indexing model from the issue text.