基于EEMD_GM_ELM模型的滑坡位移预测
Landslide displacement prediction based on EEMD一GM-ELM modelAs the landslide is affected by various complex factors, the displacement data obtained often has great uncertainty in real life, so it is difficult to predict the displacement data with only one prediction model. In consideration of the fluctuation factor rainfall, this paper first uses the Ensemble Empirical Mode Decomposition (EEMD) model to decompose the landslide displacement data into trend data and fluctuation data based on the limitations of a single
prediction model, and the two improved prediction algorithm are used to predict landslide displacement data. Through threshold optimization, the number of neurons in the hidden layer of Extreme Learning Machine (ELM) is selected to establish the correspondence between rainfall and fluctuation data. According to the haracteristics of trend data and volatility data, the gray GM (1，1) model and improved ELM model are used to predict the trend data and volatility data,respectively. Experimental results show that the model combines the advantages of the GM (1，1) model and the improved ELM model, and provides a reliable prediction model for landslide displacement based on the consideration of volatility factors.
Keywords:Displacement decomposition; GM(1，1);Improved ELM; Combination prediction;
基于EEMD_GM_ELM模型的滑坡位移预测
由于滑坡受各种复杂因素的影响，在现实生活中得到的位移数据往往具有很大的不确定性，因此仅用一个预测模型很难对位移数据进行预测。考虑到降雨的涨落因素，本文首先利用集合经验模态分解（EEMD）模型，基于单一的局限性，将滑坡位移数据分解为趋势数据和波动数据利用预测模型和两种改进的预测算法对滑坡位移数据进行预测。通过阈值优化，选取极值学习机（ELM）隐层神经元个数，建立降雨与波动数据的对应关系。根据趋势数据和波动数据的特点，分别采用灰色GM（1,1）模型和改进的ELM模型对趋势数据和波动数据进行预测。实验结果表明，该模型综合了GM（1,1）模型和改进ELM模型的优点，为考虑波动因素的滑坡位移预测提供了可靠的模型。
关键词：位移分解；GM（1,1）；改进ELM；组合预测；
页:
[1]