【摘要】 根据河北省某水厂数据,建立了基于MATLAB技术的深度极限学习机模型(简称D-ELM),并通过该模型对水质的pH值、浊度(NTU)和耗氧量3个指标进行了预测分类。通过D-ELM与极限学习机(简称ELM)和BP网络模型相互比较。研究结果表明,深度极限学习机的精度比极限学习机提高6.7%,预测时长比极限学习机缩短0.486 s。而深度极限学习机的精度比BP网络模型提高了26.7%,同时预测时长比BP网络模型缩短2.707 s。从而说明深度极限学习机对水质预测的合理性和可行性,其在自来水水质预测分类中具有更高的应用价值。更多还原

【Abstract】 According to the water quality observation data of a water plant in Hebei Province, a deep extreme learning machine(D-ELM) model was established based on MATLAB, and the water quality with three important indexes such as pH value, turbidity(NTU) and oxygen consumption was predicted and classified. The deep extreme learning machine is comparedwith the extreme learning machine and BP network. The results show that the precision of the deep extreme learning machine is 6.7% higher than that of the extreme learning machine, and the prediction time is 0.486 s shorter than that of the extreme learning machine. The precision of the deep limit learning machine is 26.7% higher than that of the BP network model, and the prediction time is 2.707 s shorter than that of the BP network model. It shows that deep extreme learning machine has reasonable feasibility for water quality prediction and classification, and it has a higher application value in water quality prediction and classification.更多还原