Methods of statistical analysis of quantitative data obtained by toxicological bioassays using rodents in Japan: historical transition of the decision tree.

Massive amounts of quantitative data are being obtained by toxicity tests with time. These data sometimes include the data with a different distribution pattern or with different numbers of animals in each group. The tree-type algorithm has long been used commonly to promptly analyze the difference between the control group and dosage groups. In Japan, the decision tree has been used since 1982 with several modifications of the initial procedure. Recently, the method has been revised to have a high power. The changes have been made in two directions; one is a sophisticated method and the other is a simplified method. In this paper, the historical transition of the decision tree is described and the future of the decision tree forecasted.