Price Prediction Analysis of Carbon Finance Market Based on QPSO-LSSVM Algorithm

Abstract

At present, the swarm intelligence optimization algorithm represented by QPSO and the LSSVM algorithm have been widely studied and applied due to their advantages of distribution, self-organization, and strong robustness. The international carbon market price presents non-linear, non-stationary, multi-frequency and other irregular characteristics. The lack of timely and effective carbon finance (CF) market price prediction is one of the serious reasons for the loss of carbon assets in my country. Therefore, it is particularly necessary and urgent to use relevant algorithms to study the price prediction of the CF market. This paper takes the prediction of the CF market price as the research object, and optimizes the selection and parameter setting of the kernel function of the QPSO-LSSVM model through the “second-order oscillation” and the introduction of a compression factor. This article introduces the price prediction process based on the QPSO-LSSVM algorithm in detail, and optimizes the QPSO-LSSVM parameters. Finally, by comparing other commonly used prediction models, the results show that the MAD of the PSO-ANN model is 0.0769, the MAPE is 0.0121, and the RMSE is 0.0965; the MAD of the QPSO-LSSVM model is 0.0657, the MAPE is 0.0104, and the RMSE is 0.0861. This shows that the prediction model proposed in this paper has the characteristics of high operating efficiency and higher prediction accuracy, and is more suitable for short-term prediction of CF market prices.