考虑数据去噪分解的滑坡位移组合预测研究

为准确掌握滑坡位移变化规律,基于滑坡变形监测结果统计,对位移数据进行去噪分解处理,将滑坡位移数据分解为趋势项和误差项,并分别利用优化多核极限学习机和Arima模型构建预测模型,以实现滑坡位移的组合预测.结果表明:Morlet复小波较传统去噪模型分解效果更优,且通过优化处理,能更好地提高其分解能力;通过对多核极限学习机的...

Study on combination prediction of landslide displacement considering data denoising decomposition

In order to accurately grasp the variation law of landslide displacement,based on the statistics of landslide deformation monitoring results,the denoising decomposition processing was conducted on the displacement data to decompose the landslide displacement data into the trend term and error term,and the prediction models were constructed by using the optimized multi-core extreme learning machine and ARIMA model respectively,so as to realize the combination prediction of landslide displacement.The results showed that the Morlet complex wavelet had better decomposition effect than the traditional denoising model,and through the optimization processing,its decomposition ability was better improved,which was suitable for the denoising processing of landslide displacement data.Through the progressive optimization processing of multi-core extreme learning machine,the prediction accuracy of trend term could be effectively improved,and through the error correction prediction of ARIMA model,the overall prediction accuracy could be further improved.The average relative error of the results were all less than 2%,which verified the applicability of the combination prediction idea in landslide displacement prediction.Through the extrapolation prediction,it was concluded that the landslide displacement would still further increase and tend to the adverse direction of development,so it was necessary to strengthen the disaster prevention and control to avoid the disaster losses.The results can provide theoretical guidance for the prevention and control of landslide disasters.