Prediction of air pollution based on FCM-HMM Multi-model

摘要

Air pollution is a growing problem arising from various activities and events. Monitoring and prediction of air pollution index remains an important public health concern. In this paper, FCM-HMM Multi-model is proposed to model the air pollution atmosphere system. First, FCM-HMM clustering is developed to mine the inherent states behind the meteorological observation sequence. Second, TS fuzzy inference multi-model is constructed for each state to predict the air pollution index. The experiment results proved the feasibility of the model on air pollution data in Beijing.