Air Quality Monitoring and Disease Prediction Using IoT and Machine Learning

Abstract

Air quality prediction focuses mainly on  these industrial areas. Industrial level usage of this project  requires expensive sensors and huge amount of power  supply. According to the World Health Organization  (WHO), major air pollutants include particulate pollution,  carbon monoxide (CO), Sulphur-di-oxide (SO2) and  nitrogen oxide (NO2). In addition to these mentioned gases,  PM or Particulate Matter and VOC or Volatile Organic  Compounds components also cause serious threats. Long  and short-term exposure to air suspended toxicants has a  different toxicological impact on humans. Some of the  diseases include asthma, bronchitis, some cardiovascular  diseases, and long-term chronic diseases such as cancer,  lung damage and in extreme cases diseases like pulmonary  fibrosis. In this proposed system, an IoT prototype of a  large-scale system which uses high-end and expensive  sensors that measures the various air pollutants in the  atmosphere is designed. Gas sensors are used in this  prototype to record the concentration of the various  pollutants that are encountered in the air on a regular basis.  The framework uses stored data to train the model using  multi-label classification with Random Forest algorithm,  XG Boost algorithm in the local system. The real time data  obtained using the different sensors is tested and the results  obtained would be used to predict the possibilities of  diseases such as asthma, lung cancer, ventricular  hypertrophy etc. and the Air Quality Index (AQI) are calculated. In addition to this, preventive suggestions are  also provided which is merely a cautionary message  displayed on our LCD display to vacuum clean the room or  mop the room thoroughly.