[关键词]
[摘要]
为了更精确地预测SO2排放质量浓度,解决非线性随机预测问题,提出了一种基于随机森林特征选择的GWONBEATS算法。通过随机森林算法筛选输入参数的特征,使用灰狼优化算法对NBEATS算法的超参数进行优化;与长短期记忆网络(Long ShortTerm Memory, LSTM)、门控循环神经网络(Gated Recurrent Unit,GRU)以及NBEATS算法对比分析,验证了GWONBEATS算法的有效性。将本算法应用于某大型电网公司大数据平台,探索了复杂智能算法在大数据平台上开展污染物排放预测的可行性。研究结果表明,相较于长短期记忆网络、门控循环神经网络和NBEATS方法,GWONBEATS算法预测误差更小, 其中平均绝对百分比误差MAPE为1.50%,相对均方误差RMSE为0.42,平均绝对误差MAE为0.33,决定系数R2为0.97。
[Key word]
[Abstract]
To predict SO2 emission mass concentration more accurately and to solve the nonlinear stochastic prediction problem, a novel grey wolf optimization (GWO) deep learning architecture NBEATS algorithm based on random forest feature selection was proposed. The features of the input parameters were screened by the random forest algorithm, and the hyperparameters of the NBEATS model were optimized using the GWO algorithm; the effectiveness of the proposed algorithm was verified by comparing it with long shortterm memory network (LSTM), gated recurrent unit (GRU) and NBEATS. The algorithm was applied to a large power grid company′s big data platform to explore the feasibility of complex intelligent algorithms to carry out pollutant emission prediction on a big data platform. The results show that the GWONBEATS algorithm has less error compared to LSTM, GRU and NBEATS methods, where MAPE is 1.50%, RMSE is 0.42, MAE is 0.33, and R2 is 0.97.
[中图分类号]
TK284
[基金项目]
国家自然科学基金资助项目(51506052)