Multifactor dynamic rough prediction models methods for complicated system

In this paper, as for multifactor prediction of complex systems, a dynamic rough prediction model and method (abbreviated to DRPM) is proposed. This method based on pattern recognition, with the tool of rough set, deals with datum, selects characteristics, reduces factors, draws the typical patterns of factors and prediction indices and the probable description of relevant relation. Thus the model is established. When new information is acquired, the prediction model is modified, so the dynamic prediction model is established which not only avoids the difficulty to set up accurate analysis mathematical models, but also considers the influence of uncertain factors. The instance shows that DRPM is simple, feasible, effective, and of high precision.