The interaction concept of K-systems theory

Klir systems theory or K-systems theory is an offshoot of reconstructability analysis. It provides a setting wherein common problems of statistics can be solved using the power of information theory. A cursory comparison of K-systems theory and the analysis of variance has been given [Jones, 1986]. However, the details of how to compute a K-systems interaction were not given, nor were any quantitative differences demonstrated, and the substate cognitive content measure suggested failed to provide an adequate representation of a multivariate interaction. This has led to some confusion and criticism of the work. This paper provides the fine details of the interaction concept in the K-systems and statistical context, as well as a comparative example which shows greatly disparate results. From the K-systems interaction defined here, we see that the conclusions of the original work remain valid. That is, K-systems constitute a correct model for interactions.