Gold Stock Price Forecasting Based on Nonlinear Weighted Particle Swarm (IPSO) Optimised Support Vector Machine (SVM) Time Series

Wang, Han and Dong, Xinqi and Qu, Haichen and Liao, Jiashu and Ma, Danqing (2024) Gold Stock Price Forecasting Based on Nonlinear Weighted Particle Swarm (IPSO) Optimised Support Vector Machine (SVM) Time Series. Advances in Economics, Management and Political Sciences, 85 (1). pp. 118-124. ISSN 2754-1169

Abstract

The price of gold, as an important precious metal, is highly volatile and uncertain as it is affected by the economic and political situation in the global market. Therefore, forecasting gold price is of great significance for investors, policy makers and economists. In this paper, an algorithm based on nonlinear weight decreasing PSO-SVR univariate time series prediction is proposed for forecasting gold price. The algorithm can help investors, policy makers and firms to understand market trends and price fluctuations and make more informed decisions. The algorithm is based on a nonlinear weighted particle swarm (IPSO) optimised support vector machine (SVM) time series model, which is trained with training set data and validated using test set data. Y-X scatter plots are plotted for the predicted and real values of the training set, and line plots of the predicted and real values of the training set are plotted in the coordinate system, and the results show that the algorithm is able to predict the price of the gold stock well, and the predicted and real values of the price of the gold stock can be very close to each other, both in the training set and the test set. The values of the model evaluation indexes R2, MAE, MBE and MAPE show that the algorithm can predict the gold stock price very well, and the prediction results in the test set do not deviate much from the training set. Therefore, the algorithm can provide useful market information and decision support for investors, policy makers and enterprises.