上海大学学报(自然科学版)

• 计算机工程与科学 • 上一篇    下一篇

基于时序挖掘的时间融合算法及在海表面温度预测中的应用

徐凌宇,方晓君,徐仁杰,沈立炜   

  1. 上海大学 计算机工程与科学学院,上海 200072
  • 收稿日期:2007-03-27 修回日期:1900-01-01 出版日期:2007-10-20 发布日期:2007-10-20
  • 通讯作者: 徐凌宇

Time-Fusion Based on Time-Series Mining and
Application of Sea Surface Temperature Forecast

XU Ling-yu,FANG Xiaojun,XU Renjie,SHEN Liwei   

  1. School of Computer Engineering and Science, Shanghai University, Shanghai 200072, China
  • Received:2007-03-27 Revised:1900-01-01 Online:2007-10-20 Published:2007-10-20
  • Contact: XU Ling-yu

摘要: 使用序列挖掘的时间融合模型研究数字海洋中海表面温度(Sea Surface Temperature,SST)预测,对基于时序的动态SST数据进行推测.提出基于平滑处理与支持度判断的抛物线回归模型方法,通过对曲线拐点判断方法的改进,改善抛物线回归模型在趋势预测方面存在的不足.试验证明此方法对发展趋势较为平稳的SST数据具有较好的预测效果.

关键词: 海表面温度 , 时间融合, 时序, 挖掘

Abstract: We use a time-series mining algorithm to fuse the dynamic sea surface temperature (SST) data in order to predict the future values and the trend. An improved method is proposed that uses the parabola regression model to forecast the future SST values based on median smooth and a threshold limit. By modification to the judgment of inflection points, the parabola regression model is improved in the trend prediction. Experiments show that the improved method has a better effect on the forecast of the slowly varying dynamic SST data.

Key words: mining, sea surface temperature, time fusion, time-series