Development Of Software Platform For Dynamic Prediction Of Air Quality

In recent years,with the environment caused by a series of problems,a better understanding of environmental trends can provide environmental protection measures for the government.The environmental pollution index is an important indicator of the degree of environmental pollution,real-time view of environmental pollution index,to predict its changing trend is the most important problem to be solved.Based on the network platform to establish air quality testing and forecasting website not only to facilitate the public to see also for researchers to provide an important scientific basis.This paper develops an Internet-based online data acquisition system.In order to obtain the data of environmental pollution index of each administrative region,we use JAVA language to write the program code,with PM2.5.in website provided by the API interface to set the collection of cities and the time required to collect data,the real-time collection of environmental pollution index.Taking into account the large amount of data transmitted and the data security requirements are higher,and ultimately in the parameters of the way to choose to use POST.In the data preprocessing process,a large number of data loss for the background algorithm to bring a lot of trouble,so the use of ARX model interpolation method for missing data interpolation.In the background,the Tornado server in Python is used to interact with the hypertext markup language.The problem that needs to be solved in the front-end and background fusion process is to calculate the data for too long and the data format conversion.In this paper,multi-threaded optimization and AJAX technology to achieve asynchronous communication data exchange,and then solve the problem of computing too long;in the data format conversion on the use of lightweight JSON data exchange format.The advantage of the software platform designed in this paper is to apply the network data mining technology to the prediction of air quality.The collected environmental pollution index is divided into two parts: historical data and forecast.Historical data is the data base of the research.In this paper,the least squares(LS)and recursive forgetting factor(RFF)are used to predict the historical data in real time.In addition,the system is capable of multi-scale prediction in time and space.Continued updating of the data will also make the platform show higher prediction accuracy.Experiments show that the software development platform of air quality forecasting established in this paper has a certain guiding effect on the users in practical application.