The discussion of these two SSD packages is very brief and gives only the barest indications of functionality. Indeed, the summaries provided here are essentially all they offer. It would be helpful to discuss what options exist, if any, for … Read More
There is hardly a mention of censoring. This is unfortunate, as censored data, especially rightcensored observations, are common in regulatory risk assessment and there is not yet a general agreement on how, or even whether, such data should be analyzed … Read More
“We strongly support model averaging as a means of (partially) resolving the omnipotent and vexing issue of distribution selection in SSD modelling.” Surely the authors mean to use the word “omnipresent.”
Benchmark datasets The authors assert that it is desirable and necessary to assemble a collection reference data sets having certified properties that can be used to evaluate SSD methodologies and software tools. It is hard to argue with this and … Read More
Numerical stability This is limited to a brief discussion of the Burr family of distributions. The brief criticism given is not helpful. The paper by Shao (2000) and the documentation for the original BurrliOX software already discusses the points made … Read More
Discussion of SSD Toolbox software. Log-transformed data can be fit by six distributions: normal, logistic, triangular, Gumbel, Weibull, and Burr. It is also interesting that four different fitting methods are available. These are maximum likelihood, moment matching, cdf linearization, and … Read More
Discussion of ssdtools software. He (sic) authors note that while it is possible to fit a single model, the emphasis in this package is on model averaging. The distributions available for this purpose are log-normal, log-logistic, and gamma by default … Read More
The authors identified four options for dealing with bimodality: (i) use all data to fit the SSD using a unimodal model (i.e. do not account for bimodality); (ii) use only the data from the most sensitive species; (iii) use all … Read More
The authors attempt to remove unintentional over-weighting of similar model by restricting candidate models to one for each “shape.” It is not clear that this is desirable or what “shape” means here. If models that have the same “shape”, whatever … Read More
The authors advocate model averaging of HC5 estimates using AICc as weight. More specifically, they use the difference, AIC-AICmin, as weight (dropping the small c from AICc for simplicity. They note some problems when there is censoring, which is a … Read More