Towards a novel approach based on machine learning to assess environmental suitability in smart campuses

Abstract

In the context of a 'smart campus', which serves as a specialized version of a smart city designed for educational institutions, there lies the potential to employ cutting-edge technologies like the Internet of Things (IoT), artificial intelligence (AI), and big data analytics[1]. These tools are directed towards establishing a more efficient, sustainable, and comfortable environment for the entire campus community. This study takes a step forward by concentrating on creating an optimal indoor environment within the campus, specifically tailored to individuals with specific environmental sensitivities. In our endeavor, we give precedence to two crucial environmental parameters: the heat index[2] and the air quality index[3], both known to have a significant impact on individual comfort and well-being. The goal is to optimize these factors, fostering a secure and favorable environment. To accomplish this, we propose the development of predictive models capable of forecasting heat index and air quality index values. By employing three prominent models - Long Short Term Memory (LSTM), Gated Recurrent Units (GRU), and one-dimensional Convolutional Neural Networks (Conv1D) - we seek to determine the suitability of an environment, ultimately enhancing the well-being of those within the campus.