基于改进的可拓神经网络的人防工程空气质量评估方法的研究

使用可拓神经网络构建人防工程空气质量评估模型,在考虑空气污染因素较多时,训练可拓神经网络的过程运算量大、收敛困难。针对这种情况,引入灰色关联聚类理论,综合考虑空气污染物之间的内在联系,提出了基于灰色关联聚类性能优化的可拓神经网络空气质量评估方法。首先,使用灰色关联聚类方法计算空气污染物之间的相关度,并进行聚类分析。然后,针对空气污染物聚类后的结果再利用可拓神经网络空气质量评估模型对人防工程空气质量进行评估。最后,对优化后的模型进行验证分析,结果表明使用灰色关联聚类对空气污染物进行聚类分析,再利用可拓神经网络对人防工程空气质量进行评估,能够保证在运算结果不失真的同时提升可拓神经网络的运算效率,为人防工程空气质量评估提供了一种新的思路。

Research on Air Quality Assessment Method of Civil Air Defense Engineering Based on Improved Extension Neural Network

Using the extension neural network to construct air quality evaluation model for civil air defense engineering,when we take more air pollution factors into account,the computation of training the extension neural network is larger and the convergence is more difficult.In view of this situation,the grey relational clustering theory is introduced,considering the intrinsic relationship between air pollutants,an air quality evaluation method based on grey correlation clustering performance optimization is proposed.Firstly,grey correlation clustering method is used to calculate the correlation degree between air pollutants and cluster analysis.Then,we can use the extension neural network air quality evaluation model to evaluate the air quality of civil air defense engineering based on the results of the air pollution cluster.Finally,by verifying and analyzing the optimized model,we can find that when we use the theory of the gray correlation clustering to cluster the air pollutants and then use the extension neural network to evaluate the air quality of civil air defense engineering,which can ensure that the calculation results are not distorted while improving the computational efficiency of the extension neural network.This method provides a new idea for the air quality assessment of civil air defense engineering.