Research On Air Quality Index Based On Fuzzy Time Series Model

In recent years,the issue of air pollution has received widespread attention from all walks of life.The air quality affects the country’s economic development as much as it affects the health of the people.Therefore,the control of air pollution is a problem we must face.The “Ten Atmospheres” and the “Blue Sky Defense Plan” issued successively by our country have shown the determination to tackle the problem of atmospheric pollution and improve air quality.2020 is the final year of my country’s three-year “Blue Sky Defense” and a “milestone year” for the coordinated control of climate change and air pollution in my country.After vigorous efforts in various regions,the overall air pollution problem in my country has been significantly improved,but this work still needs to be continued,and strive to solve the difficulties of the “smog siege” as soon as possible.It is one of the effective methods to control the future change trend of air quality by studying the changing law of the air quality index.After the unremitting efforts of many researchers,the research on air quality has achieved remarkable results,from the application of the classic time series ARIMA model to the application of support vector machines,neural networks and combination models,as well as the introduction of multiple related factors.However,researches based on time series all assume that the observed value is generated by the actual value plus random errors,and believe that the uncertainty of the data is completely described by randomness,ignoring the ambiguity of the data itself.Therefore,based on the ambiguity and uncertainty of the data,this paper tries to study the changes of air quality from the perspective of fuzzy theory and establishes a fuzzy time series model.This paper establishes a fuzzy time series model based on the daily air quality index of Nanjing area,which mainly includes four steps: the division of the universe,the fuzzy set and the fuzzy historical data,the establishment of fuzzy rules and the defuzzification forecast.Aiming at the fuzzy characteristics of the data,the fuzzy C-means clustering algorithm is introduced into the interval division process to maximize the use of information in the historical data.When establishing fuzzy rules,not only the frequency of fuzzy relations,but also the order of their appearance is included in the calculation range.Secondly,establish the representative ARIMA model.Verify the validity of the fuzzy time series model by comparing with previous research methods.According to the prediction results,it can be seen that the fuzzy time series model has obtained a relatively ideal prediction effect in predicting the air quality index,which provides a new idea and direction for it.