Air Quality and Healthy Ageing: Predictive Modeling of Pollutants Using CNN Quantum-LSTM

The concept of healthy ageing is emerging and becoming a norm to achieve a high quality of life, reducing healthcare costs and promoting longevity. Rapid growth in global population and urbanisation requires substantial efforts to ensure healthy and supportive environments to improve the quality of life, closely aligned with the principles of healthy ageing. Access to fundamental resources which include quality healthcare services, clean air, green and blue spaces plays a pivotal role in achieving this goal. Air quality, in particular, is a critical factor in achieving healthy ageing targets. However, it necessitates a global effort to develop and implement policies aimed at reducing air pollution, which has severe implications for human health including cognitive impairment and neurodegenerative diseases, while promoting healthier environments such as high quality green and blue spaces for all age groups. Such actions inevitably depend on the current status of air pollution and better predictive models to mitigate the harmful impact of emissions on planetary health and public health. In this work, we proposed a hybrid model referred as AirVCQnet, which combines the variational mode decomposition (VMD) method with a convolutional neural network (CNN) and a quantum long short-term memory (QLSTM) network for the prediction of air pollutants. The performance of the proposed model is analysed on five key pollutants including fine Particulate Matter PM2.5, Nitrogen Dioxide (NO<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub>), Ozone (O<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub>), PM10, and Sulphur Dioxide (SO<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub>), sourced from air quality monitoring station in Northern Ireland, UK. The effectiveness of the proposed model is evaluated by comparing its performance with its equivalent classical counterpart using root mean square error (RMSE), mean absolute error (MAE), and R-squared (R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>). The results demonstrate the superiority of the proposed model, achieving a performance gain of up to 14% and validating its robustness, efficiency and reliability by leveraging the advantages of quantum computation.