A Neural Network Approach for Predicting Air Quality Forecasting in Smart Cities Based on ST-LSTM Technique

The Internet of Things (IoT) has recently emerged as an intriguing field of study across numerous fields, including commerce, academia, and industry. By utilizing services and apps based on the Internet of Things, "smart cities" make urban living more environmentally friendly. Smart cities that leverage the Internet of Things have the potential to increase efficiency, participation, and knowledge among city stakeholders. Over the last several years, the quantity of data generated by smart city applications that rely on the Internet of Things has grown exponentially. This specific order is required for the execution of model training, feature extraction, and data preprocessing. As part of the preprocessing phase, data is normalized, categorical variables are encoded, and missing values are dealt with. The initial stage of this process is to identify the most crucial components of the air quality index (AQI), which represents the overall air quality, and then choose them. When training ST-LSTM models, the initial step is to retrieve relevant features. A pair of state-of-the-art methods, TCN and LSTM, were surpassed by the suggested methodology. After using the method, accuracy increased by 95.53%.