No abstract

No abstract

No abstract

SUMMARYContaminant concentration data collected at Superfund sites are typically positively skewed, and the log-normal distribution is commonly used to model such data distribution. U.S. EPA guidance documents recommend the use of H-statistics to compute the upper confidence limit (UCL) of the mean of a log-normal distribution. Recent work reported in the statistical literature has shown that the UCL calculated from the H-statistics can yield extremely high false positives. In the present article we compute the Bayesian highest posterior density (HPD) credible set of the log-normal mean. Simulated results using techniques of computational geometry are presented. Several experimental results on environmental data sets reveal that the UCL obtained by using the proposed Bayesian approach is more reasonable than those obtained by using other techniques.

In an imbalanced dataset with binary response, the percentages of successes and failures are not approximately equal. In many real world situations, majority of the observations are "normal" (i.e., success) with a much smaller fraction of failures. The overall probability of correct classification for extremely imbalanced data sets can be very high but the probability of