【摘要】 当前基于云平台下线性回归算法用于能见度、湿度与气溶胶PM₁₀之间相关性研究较少,针对现代大规模数据线性回归在单机计算时间过长的问题,设计云环境下DMLR（Distributed Multiple Linear Regression）模型用于能见度、湿度与气溶胶PM₁₀相关性的研究,实验分析表明,湿度区间一致大气气溶胶PM₁₀浓度越大能见度就越小,能见度区间一致大气气溶胶PM₁₀浓度越低湿度越大。实验结果还发现湿度介于40%-90%,能见度介于8km-19kmDMLR预测效果最好。DMLR算法模型在时间性能方面要优于传统回归模型。更多还原

【Abstract】 At present,there are few researches on the correlation between visibility,humidity and aerosol PM₁₀ based on the linear regression algorithm under the current cloud platform.For the problem that the linear computing of modern large-scale data is too long in the single machine,DMLR（Distributed Multiple Linear）is designed in the cloud environment.Regression model is used to study the correlation between visibility,humidity and aerosol PM₁₀.The experimental analysis showed that the higher the concentration of atmospheric aerosol PM₁₀ concentration,the smaller the visibility,and the lower the concentration of atmospheric aerosol PM₁₀ concentration.The experimental results also show that the humidity is between 40% and 90%,and the visibility is best when the visibility is between 8 km and 19 km.The DMLR algorithm model outperforms the traditional regression model in terms of temporal performance.更多还原