A new GM(1,1) model for predicting oscillation data sequence

The grey model GM (1,1) can be used to predict the data, however, the model usually requires the sequence monotonic. In the prediction of uniform distribution oscillation sequence, using GM (1,1) model directly is not satisfactory. Based on the feature of GM (1,1) model, this paper present an improved algorithm. In the prediction process, we generate residual sequence by raw data firstly, transform them into monotonically increasing sequence through triangle transformation secondly, and then make prediction via using GM (1,1) model, and revert data at last. An example comparison indicated that the data predicted by this method not only has smaller prediction error, but also meets the overall trend, making better prediction than the old ones. So this algorithm can be used for analysis of system reliability.