周靖楠1,刘振男1,2.基于自适应差分进化算法优化极限学习机的干旱预测方法[J].水电能源科学,2018,36(6):6-9
基于自适应差分进化算法优化极限学习机的干旱预测方法
Drought Prediction Method Based on Self adaptive Differential Evolutionary Algorithm and Extreme Learning Machine
  
DOI:
中文关键词:  极限学习机  适应差分进化算法  干旱  SSTA  预报因子
英文关键词:extreme learning machine  self adaptive differential evolutionary algorithm  drought  SSTA  forecasting factor
基金项目:贵州省科技厅基金项目(黔科合LH字\[2016\]7096)
作者单位
周靖楠1,刘振男1,2 1. 贵州理工学院贵州 贵阳 5500012. 河海大学 水文水资源学院江苏 南京 210098 
摘要点击次数: 29
全文下载次数: 
中文摘要:
      针对极限学习机在实际应用时随机选取初始权值与阈值易导致其稳定性弱及泛化能力差的问题,利用自适应差分进化算法对其进行改进,构建了自适应差分进化极限学习机预测模型,并选用海表异常温度作为该模型的输入因子,对研究区域的干旱进行预测。结果表明,以海表异常温度作为模型的输入因子,应用极限学习机能有效地进行干旱预测,通过自适应差分进化算法优化的极限学习机应用于干旱预测,其精度与稳定性均有所提高。
英文摘要:
      The initial weights and thresholds selected randomly in extreme learning machine training process are easily to result in poor stability and generalization ability. Extreme learning machine was improved by using self adaptive differential evolution algorithm, and the self adaptive differential evolution extreme learning machine model was built. The sea surface temperature anomalies (SSTA) were selected as the input factors to forecast drought in the study region. The results show that extreme learning machine can be effectively applied in drought prediction with SSTA, and the self adaptive differential evolution extreme learning machine model enhances the accuracy and stability of the drought prediction.
查看全文  查看/发表评论  下载PDF阅读器