Predicting Daily Ozone Concentration Maxima Using Fuzzy Time Series Based on Two-Stage Partition Method

Abstract

Air pollution is associated with health effects,therefore air pollution is an important and popular topic. In Taiwan, the pollutant standard index (PSI) has been adopted to assess air pollution. The report of primary focus is for health,and the current PSI sub-indices reflect the PM10 and O4 measured concentrations. Therefore, this study uses O3 attribute to evaluate air quality. This paper proposes a new fuzzy time series based on two-stage partition method to predict air quality by daily maximum O3 concentration: (1) fuzzy time series based on uniform discretion method to partition the universe of discourse; (2) fuzzy time series based on cumulative probability distribution approach to partition the universe of discourse. The proposed methods are verified by practical collected dataset. From the results, the two proposed methods both outperform the listing methods in RMSE. The fuzzy time series methods can provide more accuracy predictions for the daily maxima ozone concentration.