Consistency-based prediction using extensions of information systems —An experimental study
Krzysztof Pancerz · 2008
The paper is devoted to the application of the extensions of information systems to state prediction problems. An information system (in the Pawlakpsilas sense) can describe the states of processes observed in a given concurrent system. If we extend a given information system by adding some new states which have not been observed yet, then we are interested in the degrees of consistency (called also consistency factors) of added states with the knowledge of state coexistence included in the original information system. Such information can be helpful in predicting the possibility of appearing given states in the future in the examined system. The consistency factor computed can be between 0 and 1, 0 for the full inconsistency and 1 for the full consistency. The experiments show that the states from extensions of original information systems, having greater values of consistency factors, appear significantly more often in the future.