基于PSO优化极限学习机神经网络的空气质量预报

为了提高空气质量预测精度,提出一种基于粒子群算法优化极限学习机的空气质量预测模型.运用粒子群算法优化极限学习机的初始权值和偏置,在保证预测误差最小的情况下实现空气质量最优预测.选择平均绝对百分比误差、均方根误差和平均绝对误差作为评价指标,通过PSO-ELM、GA-ELM、SOA-ELM、DE-ELM和ELM五个模型预测结果对比发现,PSO-ELM可以有效提高空气质量预报的预测精度,可为空气质量预测提供新的方法和途径.

Air quality prediction based on PSO extreme learning machine of neural network

In order to improve the accuracy of air quality prediction, an air quality prediction model based on the particle swarm algorithm optimization for extreme learning machine(PSO-ELM)was proposed. The initial weight and bias of extreme learning machine were optimized with the particle swarm, and the optimal air quality prediction was achieved when the prediction error was minimized. By choosing mean absolute percentage, root mean square and mean absolute errors as evaluation criteria and comparing the results of five models, i.e.PSO-ELM, GA-ELM, SOA-ELM, DE-ELM and ELM, it was discovered that PSO-ELM can effectively improve the accuracy of air quality prediction and pave new ways and tracks for air quality prediction.