Contribution to Relation Analysis in Growing Databases.

For the analysis of nominally scaled data, such as those frequently occurring in industrial quality control, a method is developed which makes it possible to determine cause-effect relations even in the presence of a growing database. In this case even complex multivariate relations, such as those of exploratory data analysis, can be investigated. The method developed contains both the generation and the testing of hypotheses, so that in the end only a causality analysis of the hypothesis has to be carried out. Thus, for the first time, cause-effect relations can be analyzed without prior knoweledge even in the presence of a growing database.