In order to prevent air pollution and improve the living environment for residents, it is particularly important to carry out air quality forecasting. Air quality is affected by many factors, and showed significant non-linear features. Output–input feedback Elman (OIF Elman) neural network can effectively solve non-linear problems. However, the disadvantages of OIF Elman neural network are easy to fall into local minimum, slow convergence and inflexibility. Chicken swarm optimization (CSO) algorithm has high operating efficiency and fast convergence speed. Therefore, this paper proposes an air pollution prediction model for OIF Elman neural network based on the CSO algorithm (CSO-OIF Elman neural network model). Evaluation indicators are absolute average error and accuracy rate. The efficacy of the proposed model is compared with other models such as traditional Elman neural network model, OIF Elman neural network model and Elman neural network model based on CSO algorithm (CSO-Elman neural network model). The experimental results show that CSO-OIF Elman neural network model has the best accuracy and the smallest absolute average error value, and has higher nonlinear fitting capabilities and generalization capabilities. The establishment of this model can provide useful reference value for atmospheric prediction research.

为了防止空气污染并改善居民的生活环境，进行空气质量预测尤为重要。空气质量受许多因素影响，并表现出明显的非线性特征。输出-输入反馈Elman（OIF Elman）神经网络可以有效地解决非线性问题。但是，OIF Elman神经网络的缺点是容易陷入局部最小值，收敛缓慢和缺乏灵活性。鸡群优化算法具有较高的运行效率和较快的收敛速度。因此，本文提出了一种基于CSO算法的OIF Elman神经网络空气污染预测模型（CSO-OIF Elman神经网络模型）。评估指标是绝对平均误差和准确率。将该模型的有效性与其他模型进行比较，例如传统的Elman神经网络模型，OIF Elman神经网络模型和基于CSO算法的Elman神经网络模型（CSO-Elman神经网络模型）。实验结果表明，CSO-OIF Elman神经网络模型具有最佳的精度和最小的绝对平均误差值，并且具有较高的非线性拟合能力和泛化能力。该模型的建立可以为大气预报研究提供有用的参考价值。并且具有更高的非线性拟合能力和泛化能力。该模型的建立可以为大气预报研究提供有用的参考价值。并且具有更高的非线性拟合能力和泛化能力。该模型的建立可以为大气预报研究提供有用的参考价值。