基于神经网络模型的高架轨道噪声烦恼度预测

Prediction of Annoyance Degree of Elevated Track Noise Based on Neural Network Model

为了准确高效预测高架轨道噪声居民主观烦恼度,建立BP神经网络模型并对其进行优化,并对BP神经网络模型的训练和检验结果进行相关性分析,结果表明该模型具有很好泛化能力和学习能力。相较于多元线性回归,BP神经网络模型更适合应用于高架轨道交通噪声的烦恼度预测。将建立的烦恼度预测模型与通过仿真得到的噪声数据结合,可以为高架轨道交通噪声居民主观烦恼度的评估提供新的方法。

In order to accurately and efficiently predict the subjective annoyance of residents due to elevated track noise, the BP neural network model is established and improved, and the correlation analysis of the training and test results of the BP neural network model is carried out. The results show that the model has a good generalization ability and learning ability. Compared with multiple linear regression, the BP neural network model is more suitable for the annoyance prediction of viaduct rail traffic noise. Combining the proposed annoyance prediction model with the noise data obtained by simulation can provide more intuitive information for rail transit planning or construction, and is suitable for green transportation development planning.