Improved combined system and application to precipitation forecasting model

Abstract

Reliable precipitation forecasting is essential for effective water management and timely warning of natural disasters such as floods and droughts. However, precipitation is a nonlinear water vapor cycle with certain spatial and temporal dependence, and stable prediction accuracy cannot be obtained by using a single model. Therefore, this paper proposes a novelty prediction model based on original feature extraction and an improved multi-objective swarm intelligence optimization algorithm, and it carries out multi-step prediction tests for two sites in the arid/semi-arid region (Qilian Mountain-Hexi Corridor). Finally, through the 19 comparison models, 5 evaluation indexes and 3 model performance tests, it is confirmed that the precipitation combined forecasting model constructed in this study is a reliable prediction system with optimal parameters. And it can provide favorable technical support for weather forecasting.

Keywords

• Precipitation combined forecasting model
• Data original feature extraction
• Improved swarm intelligence optimization algorithm
• Regions of the arid/semi-arid