Empirical Discovery of Potential Value Leaks in Processes by Means of Formal Concept Analysis

Process improvement programs rely heavily on various techniques for the identification of deficiencies in the processes and for the explanation of their root causes, in order to formulate adequate remediating actions. Six-sigma and Lean Management approaches are useful statistical techniques in this context. In the practical application of such techniques, the process analyst is often restricted by the semantic expressiveness of the process analysis models. Many popular process modeling techniques also suffer from limitations in the adequate representation of probabilistic behavior in processes. The latter is important, for example, for detecting the impact of exceptions and variations in processes. In this paper additional techniques are proposed based on Formal Concept Analysis (FCA). This is primarily based on the semantic richness of FCA-schema´s, which are algebraic lattices. It is shown how many-to-many transitions in processes can easily be identified in an FCA-based Process representation. These transitions are not only the expression of potential Value Leaks as they also lead to explanations of their root causes. The techniques described in this paper are not a replacement, but rather an augmentation for Six-sigma and Lean Management approaches. In the paper several examples that are known in the Process Mining literature are presented and discussed. The techniques that are proposed are scalable and extend easily to large scale examples.