Abstract

Statisticians use the term interaction for situations whereinthe effect of one variable depends upon another variable[McNeil, Francis, McNeil, 1975]. A deep knowledge aboutinterdependencies and interactions among elements withina domain is conducive to devising action plans towards adesired situation. It allows for product or systems optimization,preventing pitfalls and diagnosing malfunctions.As part of a project to develop a Knowledge Based Systemfor supporting research activities in the field of polymericadhesives, we compared different methods of assessing interactions.Here, we employed General Liner Models as aclassical approach, Characteristic Angles, Self-OrganizingMaps (SOM) and Interactive Visual Feature ExtractionMethods (IVFEM) as alternative techniques. The latter isbased on a data projection method, where attributes aremapped to areas that can then be selected interactively. Inthis manner, we can describe interaction in terms of sharedattribute regions in proportion to combined area. As aquantization method, we initialize SOMs with characteristicangles to illustrate partial interactions. Characteristicangles based on multivariate analysis of the angles betweeneigenvectors can be utilized to describe partial interactionsassociated with an explained variability.