RAPHS : An Approach for Modeling and Comparing Graph-Based and Sequential Hypotheses

The analysis of sequential patterns is a prominent research topic. In this paper, we provide a first formalization of a graph-based approach, such that a directed weighted graph/network can be extended using a sequential state transformation function, that “interprets” the network in order to model state transition matrices. We exemplify the approach for deriving such interpretations, in order to compare these and according hypotheses in the context of an industrial application. Specifically, we present first results of applying the proposed approach for topology analysis and anomaly analytics in a large-scale sensor-network.