基于时空优化深度神经网络的AQI等级预测  

AQI levels prediction based on deep neural network with spatial and temporal optimizations

作　　者：董婷 赵俭辉 胡勇[2] 

机构地区：[1]武汉大学计算机学院软件工程国家重点实验室,武汉430072 [2]武汉大学资源与环境科学学院,武汉430079

出　　处：《计算机工程与应用》2017年第21期17-23,41,共8页Computer Engineering and Applications

基　　金：中国空间技术研究院创新基金(No.CAST2014);湖北省科技支撑计划(No.2014BAA149);中央高校基本科研业务费专项(No.2042016gf0023)

摘　　要：针对现有空气质量预测方法精度偏低、对噪声敏感等问题,提出一种基于堆栈降噪自编码(Stacked Denoising Auto-Encoders,SDAE)模型的空气质量等级预测方法。首先以武汉市历史空气质量和气象监测数据为研究对象,建立SDAE模型逐层学习原始数据的特征表达,并将最后一层特征与分类器连接完成预测模型的调优。同时改进多参数网格搜索法,选取了最优的超参数组合。然后在测试集上进行预测,并用预测值与实际值之间的平均绝对误差和均方误差等指标作为预测性能评价标准。通过与其他网络模型的实验对比,证明了SDAE模型对于空气质量等级具有较优的预测性能。最后从时间、空间、时空三个角度对该模型输入进行优化,实验结果表明基于空间优化的SDAE模型预测性能提升最为明显,能够得到比传统方法更加精确的预测结果。The existing air quality prediction models have lower precision, and sensitive to noisy data. Thus a new method is proposed for AQI levels prediction based on Stacked Denoising Auto-Encoders（SDAE）model. Firstly, the historical air quality and meteorological monitoring data of Wuhan city are taken as research object. SDAE model is established to study the characteristic expression of the original data layer by layer, and the last layer is connected with a classifier to tune the prediction model. The optimal set of hyper-parameters is found through improved grid search algorithm for multiparameters. Then, the prediction is obtained from the test set. The indicators such as mean absolute error and mean square error between the predicted value and related actual value are used as the evaluation standards for forecasting performance. Compared with other network models, it can be proved that SDAE model has better predictive performance. Finally,the input data is optimized considering their spatial and temporal relations. Experimental results show that the spatial optimization based SDAE has the most improvement for predictive performance, and it can obtain more accurate predictions compared with the traditional methods.

关 键 词：AQI等级 预测 堆栈降噪自编码 优化 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]