摘要
针对银行间质押式回购利率受到多种外部因素影响较难预测的问题,本研究提出基于TEI@I思想的EEMD-XGBoost-ARIMA的银行间质押式回购利率预测算法,构建基于宏观经济、货币政策、流动性因素、相关利率指标等多维度的预测指标体系,引入百度搜索指数作为文本指标进一步提升预测精度。结果表明,基于TEI@I的组合预测模型表现优于传统时间序列预测模型ARIMA,对银行间质押式回购利率的有效预测可为金融机构、投资者和相关监管部门进行流动性管理提供有效依据。
To better predict interbank pledged repo rates due to the influence of multiple external factors,a TEI@I-based EEMD-XGBoost-ARIMA prediction algorithm was proposed.This algorithm constructed a multi-dimensional forecasting index system based on macroeconomic,monetary policy,liquidity factors,relevant interest rate indicators,etc.The prediction accuracy is further enhanced by the introduction of Baidu search index as a textual indicator.The results showed that the combined forecasting model based on TEI@I outperformed the traditional time series forecasting model ARIMA,and the effective forecasting of inter-bank pledged repo rates can provide an effective reference for financial institutions,investors and relevant regulators in liquidity management.
作者
韩子哲
周子杰
张冉
周旻
李秀婷
李想
陈星
王雨楠
毛贯中
左光远
谢坤
王莹莹
HAN Zizhe;ZHOU Zijie;ZHANG Ran;ZHOU Min;LI Xiuting;LI Xiang;CHEN Xing;WANG Yunan;MAOGuanzhong;ZUO Guangyuan;XIE Kun;WANG Yingying(School of Economics and Management,University of Chinese Academy of Sciences,Beijing 100190;School of Software,Tsinghua University,Beijing 100084;CHINABOND FINANCE AND INFORMATION TECHNOLOGY CO.,LTD.Beijing 100044)
出处
《科技促进发展》
2023年第4期229-238,共10页
Science & Technology for Development
基金
2019年国家自然科学基金应急管理重点项目(71850014):房地产市场与金融风险防范,负责人:董纪昌
2020年国家自然科学基金面上项目(71974180):人口老龄化、住房租购决策与家庭金融资产配置,负责人:李秀婷
2020年中科院学部工作局中国科学院学部工作局项目(2020-ZW10-A-022):保障金融安全的科技支撑能力与对策研究,负责人:董纪昌