


摘 ?要: 時序模型作為一種預測方法,在貨運量預測、機場客流量預測、疾病發病率預測、空氣質量預測等許多重要的領域具有廣泛的應用。本文利用大同市2016年1月到2019年8月共44個月的空氣質量綜合指數數據樣本,使用牛頓插值進行了缺失值插補,根據給定的數據序列進行了時序圖、自相關圖和偏自相關圖的構建。然后,進行單位根檢驗,判斷出序列為平穩非白噪聲序列。本文使用相對最優模型識別方法確立模型的p、q值,最終建立ARIMA(2,0,1)模型,對2019年9-12月的空氣質量綜合指數進行預測。通過對模型的分析,判斷預測值比較準確。
關鍵詞: ARIMA;時序分析;非白噪聲序列;平穩序列
中圖分類號: TP391 ? ?文獻標識碼: A ? ?DOI:10.3969/j.issn.1003-6970.2019.12.020
本文著錄格式:張葉娥,高云. 基于ARIMA模型的大同市空氣質量預測研究[J]. 軟件,2019,40(12):8589
Research on Air Quality Prediction in DaTong Based on ARIMA Model
ZHANG Ye-e, GAO Yun
(School of Computer and Network Engineering, Shanxi Datong University, Datong,Shanxi 037009, China)
【Abstract】: As a forecasting method, time series model has been widely used in many important fields,such as cargo volume prediction, airport passenger flow prediction, disease incidence prediction, and air quality prediction. In this paper, 44 months of air quality composite index data samples from January 2016 to August 2019 in datong city were used to carry out missing value interpolation with Newton interpolation, and time sequence, autocorrelation and partial autocorrelation were constructed according to the given data sequence. Then, the unit root test is carried out and the sequence is determined to be a stationary non-white noise sequence. In this paper,The relative optimal model identification method was used to establish the p and q values of the model, and finally the ARIMA(2,0,1) model was established to predict the air quality index from September to December 2019. Through the analysis of the model, the prediction value is more accurate.
【Key words】: ARIMA; Time series analysis; Non-White noise sequence; Stationary series
0 ?引言
城市的空氣質量問題與居民的健康、生活、交通等問題息息相關,因此城市空氣質量成為城市居民關注的焦點問題。隨著大同市經濟的高速發展,城市人口、交通工具、石化能源消費劇增,由此帶來的污染問題對城市環境的影響倍受關注。完善大同市城市空氣質量檢測體系,對城市空氣質量檢測歷史數據的變化進行分析,預測未來的空氣質量,對即將可能出現的空氣質量問題進行預防和治理,對城市環境建設具有非常重要的意義。
1 ?預測方法
目前,常用的預測方法有BP神經網絡和時間序列分析法,邱晨等人研究了基于BP神經網絡的空氣質量模型分類預測,建立了三層神經網絡數學模型,對空氣質量等級進行分類預測,預測準確率達到90%[1]。……