On the basis of weighted non-equidistance GM(1,1) and line regression theories, we combined a weighted non-equidistance GM(1,1) model with line regression theory organically, and propose the gray linear weighted non-equidistance GM(1,1) model. Then the optimization method of the gray index v and the value of parameter m, which are vital to the model prediction accuracy, is given. In comparison with the weighted non-equidistance GM(1,1) and line regression models, the gray linear weighted non-equidistance GM(1,1) model has advantages, such as higher prediction accuracy, more valid prediction time, and more stable prediction ability. When v and m are optimized, the applicability and stability of the gray linear weighted non-equidistance GM(1,1) model is further improved."/>
灰线性加权非等距GM(1,1)形变预测模型
Abstract：On the basis of weighted non-equidistance GM(1,1) and line regression theories, we combined a weighted non-equidistance GM(1,1) model with line regression theory organically, and propose the gray linear weighted non-equidistance GM(1,1) model. Then the optimization method of the gray index v and the value of parameter m, which are vital to the model prediction accuracy, is given. In comparison with the weighted non-equidistance GM(1,1) and line regression models, the gray linear weighted non-equidistance GM(1,1) model has advantages, such as higher prediction accuracy, more valid prediction time, and more stable prediction ability. When v and m are optimized, the applicability and stability of the gray linear weighted non-equidistance GM(1,1) model is further improved.
