OSA-based interpretative structural modeling
Changzheng He · Systems engineering and electronics · 2005
The successful application of interpretative structural modeling(ISM) depends on accurate adjacency matrix of system variables. A new method called OSA (objective system analysis)-based ISM is proposed for finding the accurate adjacency matrix objectively. OSA is a self-organizing approach based on sorting-out of gradually complicated models and evaluation of them by external criterion on separate part of data sample. The minimum value of the system external criterion corresponds to the optimal system equation, which shows quantitatively whether and how the system variables interact with each other. Accurate adjacency matrix can be obtained objectively by the coefficients in the OSA models, e.g. by whether they are zero, or by a given threshold value. Then the ISM model can be built up, which can be employed to identify the system structure from the an overall aspect and extract knowledge effectively.