This research aims to improving the adaptability to different feature sequences,avoiding the"non-homology"between parameter estimation and parameter application of the model,and achieving the consistency be-tween the two criteria of model optimization and model testing.The researchers constructed a novel approximate non-homogeneous grey forecasting model with opposite-direction accumulated generation based on background val-ue,modeling mechanism and initial condition.The result shows that the novel model has high accuracy for deca-ying sequence,growing sequence,homogeneous and non-homogeneous exponential sequences,which is an effec-tive supplement to the existing grey forecasting model.
维普
万方数据