ARIMA-LP: A Hybrid Model for Air Pollution Forecasting with Uncertainty Data

Public health is seriously threatened by air pollution. Systems for early warning are crucial for preventing its negative impacts on humans. But predicting air quality is difficult because it needs precise data from time series gathered at air monitoring sites. Due to things like measurement mistakes, this data may be ambiguous. Using fuzzy symmetry triangular fuzzy numbers, we suggest a strategy in this study for preparing data that contains uncertain information. Linear programming is utilized to obtain the midpoint value for defuzzification. The fuzzy pre-processed data is then used to build a predictive model using ARIMA. Our findings imply that the linear program can greatly lower prediction errors, resulting in more precise predictions.