Feature Variables Based Predicting Modelfor Complex Production Process

A feature variables based predicting model for complex production process waspresented. Comparing with traditional methods，it bypasses the mechanism analysis，using time seriesanalysis technology and feature variables，selected from original dynamicaldatad which can reflect theoperating mode of complex production process build a new kind of predicting model.The neuralnetwork realization of this model is discussed.The results obtained by simulation show the feasibleof this method.