基于BP神经网络的空气质量模型分类预测研究  

Classification and Prediction of Air Quality Model Based on BP Neural Network

作　　者：邱晨 罗璟[1] 赵朝文 崔凯辉 QIU Chen;LUO Jing;ZHAO Chao-wen;CUI Kai-hui(Institute of Mechanical and Electrical Engineering, Kunming University of Science and Technology, Kunming 650500, China;Institute of Metallurgy and Energy Engineering, Kunming University of Science and Technology, Kunming 650093, China)

机构地区：[1]昆明理工大学机电工程学院,云南昆明650500 [2]昆明理工大学冶金与能源工程学院,云南昆明650093

出　　处：《软件》2019年第2期129-132,共4页Software

摘　　要：随着工业社会的发展,空气质量问题已经成为环保任务的主要焦点。BP神经网络作为深度学习的一种,已经在大部分领域被广泛使用。为了让广大市民更好的了解空气质量情况,本文以云南省昆明市为例,收集当地近6年的空气质量数据,并基于Python语言,在Anaconda环境下的Numpy包建立了三层神经网络数学模型,对空气质量等级进行分类预测。通过训练样本对神经网络模型的训练以及相关参数的调试,得到较好的分类预测模型。将分类结果与实际结果进行比较,结果显示,本次的神经网络模型的分类预测准确率达到90%,能够较好的分析空气质量,达到预期需求。With the development of industrial economy, air quality issues have become the main focus of environmental protection tasks. As a kind of deep learning, BP neural network has been used widely in most fields. In order to let the general public understand the air quality situation better, this paper takes Kunming City, Yunnan Province as an example, collects the local air quality data for the past 6 years, and builds a three-layer neural network mathematics model based on the Python language in the Numpy package under the Anaconda environment to classify and predict the air quality levels. Through the training sample training of the neural network model and the debugging of related parameters, a better classification prediction model is obtained. Comparing the classification results with the actual data, the results show that the classification prediction accuracy of this neural network model reaches 90%, which can better analyze the air quality and meet the expected demand.

关 键 词：神经网络 深度学习 空气质量 分类预测 

分 类 号：TP389.1[自动化与计算机技术—计算机系统结构]