“…For example, suppose a toxicity database exists on a series of N chemical agents, and it is of interest to study the characteristics of Q ≥ 1 putative potency estimators. Suppose also that the database already contains summary QSAR values for each chemical, taken from a known source such as Klopman's (1985) Computer-automated Structure Evaluation (CASE) technique. Viewing these summary values as pairs (x i , y qi ) = (QSAR i , potency qi ), q = 1, .…”