A novel application of fuzzy inference system optimized with particle swarm optimization and genetic algorithm for PM10 prediction.

Indoor air quality (IAQ) is a significant concern for occupant comfort, productivity levels, health, and well-being. The presence of harmful chemicals, inhalable particles, and toxic materials can be ten times worse as compared to outdoor air pollution. The increased concentration of PM10 pollutants in the indoor environment is a serious threat to human health in terms of critical respiratory health issues. Literature supports the application of artificial intelligence for the prediction of pollutant concentrations in building environments. In this paper, we proposed a novel application of fuzzy inference system (FIS) optimized using particle swarm optimization (PSO) and genetic algorithm (GA) for forecasting PM10 concentration. The study was conducted on experimental data obtained from an Internet of Things-based IAQ monitoring system. The real-time monitoring system measures four indoor air pollutants: PM10, PM2.5, CO2, and VOC, along with two crucial thermal comfort parameters (Temperature and Humidity). The forecasting performance of the FIS with normalized dataset was measured in terms of MSE = 4.3656; MAE = 1.9351; MAPE = 9.633% and RMSE = 2.0894. Two optimization algorithms have improved RMSE FIS-PSO = 1.0746; and FIS-GA = 0.998. The optimized FIS can be used to develop real-time air quality prediction systems for indoor environments to promote health and well-being for building occupants. [ABSTRACT FROM AUTHOR]