An intelligent fuzzy and IoT-aware air quality prediction and monitoring system using CRF and Bi-LSTM

Abstract

Recently, air pollution has been increasing drastically in the majority of metropolitan cities around the world. This is necessary to reduce air pollution, and we propose a new air quality prediction system to predict air quality and pollution levels in different seasons in Beijing, China. The proposed air quality model applies a preliminary data preprocess to get exact data, a newly proposed conditional random field (CRF), and a fuzzy rule-based data grouping algorithm (CRF-FRDGA) to group the data according to the different seasonal data by applying the necessary rules, the standard bidirectional LSTM (Bi-LSTM) for performing an effective classification and prediction process. The PM2.5 concentration in Beijing, China, is forecasted season-wise for the next five years. Various experiments have been done to prove the capability of the proposed air quality prediction system and proved better than the existing works in prediction accuracy.