An innovative hybrid model based on outlier detection and correction algorithm and heuristic intelligent optimization algorithm for daily air quality index forecasting

Air pollution forecasting plays an important role in helping reduce air pollutant emission and guiding people's daily activities and warning the public in advance. Nevertheless, previous articles still have many shortcomings, such as ignoring the importance of outlier point detection and correction of original time series, and random initial parameters of models, and so on. A new hybrid model using outlier detection and correction algorithm and heuristic intelligent optimization algorithm is proposed in this study to address the above mentioned problems. First, data preprocessing algorithms are conducted to detect and correct outliers, excavate the main characteristics of the original time series; second, a widely used heuristic intelligent optimization algorithm is adopted to optimize the parameters of extreme learning machine to obtain the forecasting results of each subseries with improvement in accuracy; finally, experimental results and analysis show that the presented hybrid model provides accurate prediction, outperforming other comparison models, which emphasize the importance of outlier point detection and correction and optimization parameters of models, it also give a new feasible method for air pollution prediction, and contribute to make effective plans for air pollutant emissions. •An innovative hybrid model is proposed for daily air quality forecasting.•Outlier point detection and correction algorithm is adopted in this study.•Hypothesis testing is used to evaluate the forecasting efficiency.•The proposed hybrid model has higher prediction level than other models.