Earthquake prediction based on time series decomposition and deep learningOA
Earthquakes are extremely destructive natural disasters,and accurately forecasting earthquakes is of great sig-nificance in reducing losses caused by earthquakes.To improve the accuracy of earthquake forecasting,we propose a novel method that integrates an improved artificial rabbit optimization algorithm(IARO),variational mode decomposition(VMD),and deep learning,named IARO-VMD-DFGNet,for earthquake time series fore-casting.Facing the challenges caused by the need for manual setting of VMD parameters,we propose the IARO to optimize the parameters of VMD,thus avoiding errors caused by manual parameter setting.Additionally,we also propose a data-driven deep learning model,DFGNet,for forecasting decomposed data.The data used in this study are earthquake catalogs,which include five variables:timestamp,longitude,latitude,depth,and magni-tude.Each variable is independently decomposed and forecasted.The performance of the model is evaluated using four metrics:mean squared error,mean absolute error,relative standard error,and root mean square error.Experimental results from four different earthquake catalogs demonstrate that the proposed model outperforms several other popular time series forecasting models.It achieves average reductions of 24.6%,18.5%,15.2%,and 13.7%across the four evaluation metrics,demonstrating significant competitive advantages.Therefore,applying IARO-VMD-DFGNet to earthquake forecasting is of great significance in reducing the harm caused by earthquakes.
Hao Luo;Zhongyi Liu;Liang Wang;Hongxing Chen;Caiyun Xia
Faculty of Information,Liaoning University,Shenyang,110036,ChinaFaculty of Information,Liaoning University,Shenyang,110036,ChinaLiaoning Earthquake Agency,Shenyang,110031,ChinaShandong Keyue Technology Co,Jinan,250000,ChinaLiaoning Earthquake Agency,Shenyang,110031,China
天文与地球科学
Artificial rabbit optimization algorithmEarthquake predictionTime seriesVariational mode decomposition
《Geohazard Mechanics》 2026 (2)
P.155-169,15
funded by the National Key R&D Program of China(No.2024YFC3013903)the Liaoning Revitalization Talents Program Outstanding Young Talent(No.XLYC2403133)the National Science Foundation of China(No.52427805)the Liaoning Pro-vincial Department of Education''s Science and Technology Research Project(No.LJKMZ20220450).
评论