Improvement of Gm (1,1) Model Based on Particle Swarm Optimization Algorithm--Model Performance Analysis Based on Epu Data
Download as PDF
DOI: 10.25236/iiiece.2022.016
Corresponding Author
Jiwei Liu
Abstract
Based on the analysis of GM (1,1) model, through theoretical derivation: (1) the first data has no effect on the predicted value, (2) when the same increment is added to each data of the original sequence, the predicted value will change. Therefore, it is proposed to insert new data before the first data of the sequence, add the same increment at each position, and use particle swarm optimization algorithm to get the best increment. By comparing the errors, this method can improve the prediction accuracy of the model.
Keywords
Gm(1, 1), Particle swarm optimization algorithm, Prediction accuracy