基于气象特征数据智能识别与评估的雷电灾害预测模型设计

Design of lightning disaster prediction model based on intelligent identification and evaluation of meteorological characteristic data

针对大部分预测模型未全面考虑雷电的时空特征而导致预测精度较低的问题,文中设计了一种基于气象特征数据智能识别与评估的雷电灾害预测模型。该模型利用数据挖掘方法获取气象特征,并计算其皮尔森相关性系数,从而确定了与雷电相关的气象特征。通过将雷电特征输入由人工鱼群算法训练后的Elman网络进行学习分析,即可得到雷电灾害的预测结果。基于Python平台对所提模型进行实验分析与测试,结果表明,该模型预测结果的平均绝对百分比误差小于0.03,且在不同时间片和空间内均具有良好的预测效果。

Aiming at the problem that most prediction models do not fully consider the space⁃time characteristics of lightning,resulting in low prediction accuracy,this paper designs a lightning disaster prediction model based on intelligent recognition and evaluation of meteorological characteristic data.The model uses data mining method to obtain meteorological characteristics,and calculates its Pearson correlation coefficient,so as to determine the meteorological characteristics related to lightning.Input the lightning characteristics into the Elman network trained by the artificial fish swarm algorithm for learning and analysis,and get the prediction results of lightning disasters.Based on the Python platform,this paper makes an experimental analysis of the proposed model,and the test results show that the average absolute percentage error of the prediction results is less than 0.03,which has good prediction results in different time slices and spaces.