Research On Air Pollution Prediction In Small Area

With the rapid development of economy,the problem of environmental pollution is becoming a serious challenge,such as acid rain,now of PM2.5.These environment problems seriously interfere with our daily life,while people pay more attention to the environmental conditions,especially for air quality.In order to control the air pollution effectively,we should make efficient monitoring and prediction.This thesis focuses on the prediction of air pollutant index in small area.Using neural network algorithm,data analysis and processing,data visualization technology,cloud storage and Python language,a concise and practicable air pollutant prediction system was designed and implemented.The designed air quality prediction system of small area include data(acquisition)storage,data analysis,data prediction and data presentation.A Python development environment is built for implementation of algorithms and basic functions.A MySQL database is set up to store the measured data,process and forecast data.All data is stored and invoked in the cloud.A prediction model is constructed to predict the development trend of pollutants,in the mean while,system provides two evaluation mode:the regional overall evaluation and regional random point evaluation.Through the integration of database,prediction model and data visualization,Python's Tkinter library is performed to write an interactive software to solve problems such as finding and displaying.Interactive software realizes the data access,analysis and visualization.By applying the system to small regions of pollutants,the forecast effect is good.The difference between the predicted result and the measured value is less than 10%,while the best can reach 3%.The designed system presents reference value to air pollution analysis and prediction.