耦合遗传算法与RBF神经网络的PM2.5浓度预测模型

基于北京市空气质量监测点获取的空气污染物浓度数据，通过遗传算法搜索径向基人工神经网络的最优隐含层神经元数目和扩展常数，构建了耦合径向基人工神经网络算法与遗传算法的预测模型，预测北京市未来一天24h平均PM_2.5质量浓度.结果表明，预测精度与泛化性能良好.该模型不需要输入气象和地理位置信息等数据，具有依赖变量少、预测精度高（R²达0.75）和运算效率高等特征，并可以通过训练样本的驱动，使自身不断优化调整.该模型预测效果可以通过扩展输入特征、增加训练样本量等方法进一步提升，可对多种时空情境下的城市空气污染进行高效率且精确的预测.

A coupling model of genetic algorithm and RBF neural network for the prediction of PM2.5 concentration

We developed a coupling model combining the radial basis function (RBF) artificial neural network and the genetic algorithm to predict the average PM_2.5 concentrations in Beijing in the next 24hours. This model mainly used air pollutant concentration data obtained by air quality monitoring stations as inputs, and relied on the genetic algorithm to determine parameters such as the number of hidden layer neurons and the spread constant. The model had a good prediction performance (R-square up to 0.75) with less data inputs because it does not need meteorological or geographical information for its training process. Further improvements can be made by using multi-source data and increasing sample size in the training process to enhance the accuracy and robustness of the model for the prediction of air pollution in different situations.