ISSN 1000-1239 CN 11-1777/TP

• Paper • Previous Articles     Next Articles

An Integrated Spatio-Temporal Forecasting Approach Based on Data Fusion and Method Fusion

Xu Wei1, Huang Houkuan1, and Wang Yingjie2   

  1. 1(School of Computer and Information Technology, Beijing Jiaotong University, Beijing 100044) 2(Institute of Computing Technology,China Academy of Railway Sciences, Beijing 100081)
  • Online:2005-07-15

Abstract: Spatio-temporal data mining is an important research topic in data mining, and in which spatio-temporal forecasting is the most widely used. In order to overcome the limitations of current spatio-temporal forecasting methods, this paper proposes a spatio-temporal forecasting approach based on data fusion and method fusion. The approach first forecasts time sequence of the target object itself with statistical principles and computes influences of neighboring objects employing neural network technique, then forecasts mixed data sequence using spatio-temporal auto-regressive model, and finally integrates the individual time sequence forecast, spatial forecast and spatio-temporal forecast through linear regression to deliver the final result. The approach was successfully used in the forecasting of railway passenger flow in an attempt to overcome the limitations of traditional railway passenger flow forecasting methods. The experimental result shows the effectiveness of the approach.

Key words: spatio-temporal forecasting, spatio-temporal data mining, fusion, passenger flow forecasting