Title :
Forecasting the air quality using OWA based time series model
Author :
Huang, Sue-Fen ; Cheng, Ching-Hsue
Author_Institution :
Inf. Manage. Dept., Univ. of Nat. Yunlin Sci. & Technol., Yunlin
Abstract :
The environmental protection conception increasing, the prediction of air quality is more and more important. The main Pollutant Standards Index (PSI) includes PM10, SO2, NO2, CO and O3 etc... The PSI will be produced and changed when combining in the air. Due to the concentrations of CO, SO2, NO2, and PM10 have declined, the focus of health studies and control efforts has increasingly turned to PM10 and O3 as the most important air pollutant species of concern. Correspondingly, the primary focus on the current understanding of the health is affected by PM10 and O3 in the Taiwan. Therefore, this study uses O3 attribute to evaluate air quality. This paper proposes an OWA based time series model to predict the air quality. Due to O3 data is belong to time series pattern, and OWA operator can aggregate multiple lag periods into single aggregated value by different situation parameters alpha. Based on the advantages of TSM and OWA, the OWA based time series model can efficiently and accurately predict PSI. In verification, this paper collects a practical data to verify the proposed method. The dataset contains records of 1061 days with O3 attribute from air qualities inspection station in Hsinchu city, Taiwan. From the results, the proposed method outperforms the listing methods.
Keywords :
air pollution measurement; forecasting theory; time series; OWA; air pollutant species; air quality forecasting; environmental protection; ordered weighted averaging; pollutant standards index; time series model; Aggregates; Air pollution; Cities and towns; Cybernetics; Humans; Machine learning; Open wireless architecture; Predictive models; Protection; Time series analysis; Air Quality; ordered weighted averaging; pollutant standards Index; time series method;
Conference_Titel :
Machine Learning and Cybernetics, 2008 International Conference on
Conference_Location :
Kunming
Print_ISBN :
978-1-4244-2095-7
Electronic_ISBN :
978-1-4244-2096-4
DOI :
10.1109/ICMLC.2008.4620967