基于改进神经网络算法的PM2.5污染信号分析检测

当前为了保证污染信号分析的精度,在对PM2.5污染进行检测的过程中,需处理的数据量过大,导致经典神经网络方法遇到矛盾数据时,需要花费大量的数据校验时间,收敛速度下降,检测效率大幅降低,提出一种基于改进神经网络算法的PM2.5污染检测方法,在分析标准神经网络算法的基础上,允许信号跳变精确度范围内,在层与层之间引入容错性变量,同时在计算阈值的过程中融入松弛变量,提高收敛速度；避免神经网络陷入局部最优解。采用改进神经网络算法,通过不断调整网络的权值以及污染阈值,对PM2.5污染信号进行高效检测。以飞利浦公司的新一代检测系统为测试器材,测试结果表明,采用所提方法得到的PM2.5污染检测效率明显提高。

PM2.5 Pollution Detection Signal Analysis Based on Improved Neural Network Algorithm

Current in order to ensure the accuracy of pollution of the signal analysis, in the testing process of PM2.5 pollution, need to deal with the amount of data is too large, lead to classical neural network method when you meet the contradiction between data need to spend a large amount of data checking time, convergence rate fell, the detection efficiency is greatly reduced, put forward a kind of PM2.5 pollution detection method based on improved neural network algorithm, based on the analysis of the standard neural network algorithm, and allow the signal jump range, precision in the introduction of fault tolerance between layer and layer variables, at the same time in the process of calculation threshold into the slack variables, improve convergence rate; Avoid neural network into a local optimal solution. With the improved neural network algorithm, and through continuous adjust the network weights and threshold, the pollution of PM2.5 pollution signal detection efficiently. In the company of a new generation of test system for test equipment, test results show that the proposed method of PM2.5 pollution detection efficiency has improved significantly.