厦门大学龚旭学术报告

发布时间:2021年07月06日 作者:张宏伟   阅读次数:[]

报告题目:A new hybrid deep learning model for forecasting oil prices

报告人:龚旭(厦门大学,副教授,研究生导师)

报告时间:2021年7月7日上午10:30-12:00

报告地点:数学楼145报告厅

报告简介: This paper proposes a novel hybrid deep learning forecasting model named Mod-EMD-LSTM based on the empirical mode decomposition (EMD) and long short-term memory (LSTM) algorithms, which consists of a Prevention module and a Prediction module. Then, taking the West Texas Intermediate (WTI) futures price as an example, several empirical studies and statistical evaluations are carried out to compare the forecasting performance of the proposed model with 15 benchmark models. The results show the excellent forecasting accuracy of the proposed model. Compared with the LSTM (EMD-LSTM) model, the R2 and directional accuracy (DA) for Mod-EMD-LSTM are increased by 19.67% and 48.649 pct, respectively, meanwhile, mean squared error (MSE) and mean absolute error (MAE) reduced by 39.26% and 28.70%, respectively. And all the above evaluation values can pass the corresponding statistical tests. Finally, through two robustness tests, on the one hand, we rule out the possibility that the full-sample decomposition of EMD may contain out-of-sample extreme values features on this occasion. On the other hand, we confirm that our model has strong robustness and generalization ability.

报告人简介:

龚旭,湖南宁乡人,厦门大学管理学院中国能源政策研究院副教授,硕士研究生导师,主要研究方向为能源金融、能源安全与金融风险管理。龚旭于2016年在中南大学获得管理学博士学位(管理科学与工程专业),于2016-2018年在厦门大学从事博士后研究工作。主持国家自然科学基金面上/青年项目、中国博士后基金特别资助/面上项目和福建省社会科学规划基金青年项目等科研项目7项。目前,已经在《Energy Economics》、《Journal of Futures Markets》、《Applied Energy》、《管理科学学报》、《系统工程理论与实践》和《中国管理科学》等国内外重要期刊发表(含录用)论文40余篇,其中被SSCI/SCI检索的论文32篇,国家自然科学基金委管理科学部指定A类重要期刊论文8篇,现/曾有ESI高被引论文10篇、ESI热点论文2篇。在攻读硕士和博士学位期间,曾获湖南省自然科学奖二等奖(2016)、中南大学十佳优秀博士研究生(2015)、 硕士/博士研究生国家奖学金(2012,2014,2015)、中南大学拔尖创新博士奖学金(2015)等奖励或荣誉称号。主要学术兼职有:中国运筹学会决策科学分会理事,《系统工程理论与实践》、《中国管理科学》、《Energy Journal》、《Energy Economics》、《Journal of Economic Dynamics and Control》和《International Journal of Forecasting》等国内外重要期刊的匿名审稿人。



打印】【收藏】 【关闭