2010 年 21 巻 1 号 p. 3-6
In recent years, it has become possible to solve complex problems in substantial science as probability models without undue simplification, thanks to the development of various methods in the field of statistics. Statistics software is now widely available that allows users to input data and obtain results automatically without intricate familiarity with the details of these methods or algorithm. It is still the responsibility of the researchers, however, to select appropriate probability models (analytic methods) to apply for the relevant problems and to discriminate between useful and unnecessary information. This paper is on the selection of statistic analytical methods based on the lecture delivered at the 39th conference.