...
  • 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Technological forecasting and social change >A novel multiscale forecasting model for crude oil price time series
【24h】

A novel multiscale forecasting model for crude oil price time series

机译：原油价格时间序列的新型多尺度预测模型

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Forecasting crude oil prices is an essential research field in the international bulk commodities market. However, price movements present more complex nonlinear behavior due to an increasingly diverse range of risk factors. To achieve better accuracy, this study explores a novel multiscale hybrid paradigm to estimate crude oil prices. The method takes advantage of the variational mode decomposition method to decompose the crude oil price into several simple models, which can be explained using regular factors, irregular factors and trends. Data characteristic analysis is conducted to identify the complexity of different components of the time series. It is important for a multiscale model to select an appropriate model to produce the optimal forecasts. Thus, the final forecasted values are generated by reconstituting all these forecasting items. By investigating the West Texas Intermediate and Brent crude oil prices, this paper presents how data characteristic identification and analysis are conducted in a multiscale paradigm. The empirical analysis proves that the proposed model can achieve superior forecasting results, which indicates the effectiveness of the multiscale model at forecasting complex time series, especially crude oil prices.
机译：预测原油价格是国际散装商品市场的重要研究领域。然而，由于越来越多的风险因素，价格变动呈现出更复杂的非线性行为。为了实现更好的准确性，这项研究探讨了一种新的多尺度混合范例来估算原油价格。该方法利用了变分模式分解方法，将原油价格分解为几种简单模型，可以使用常规因素，不规则因素和趋势来解释。进行数据特性分析以确定时间序列不同组件的复杂性。对于多尺度模型来选择适当的模型来产生最佳预测是重要的。因此，通过重构所有这些预测项目来生成最终预测值。通过调查西德克萨斯中间人和布伦特原油价格，本文介绍了数据特征鉴定和分析在多尺度范例中。实证分析证明，拟议的模型可以实现卓越的预测结果，这表明多尺度模型在预测复杂时间序列中的有效性，尤其是原油价格。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号