An approach for classification of health risks based on air quality levels

There is a need to exploit the available data collected on environment for the development of smart cities in order to improve the air quality which in turn can improve the quality of life for a city. Air pollution is becoming a serious concern to the society as the air pollutants are very hazardous in nature. Pollutants affect the health and causes respiratory and cardiac problems. If air pollutants cross the limits it might be life threatening. Software or tool can be developed whose results can be applied for predicting air pollution levels, predicting air pollution related health concerns, monitoring controlling air pollution and this is challenging one. This paper focuses on analysis of air based on the available data of various air pollutants such as NO2, SO2, CO and O3. The dataset is downloaded from Kaggle website which contains air pollutant's with corresponding AQI values. This paper implements Naive Bayes and Decision tree J48 algorithm for predicting the health concern. The categories based on Air Quality Index(AQI) are good, moderate, (unhealthy for sensitive groups) unhealthy_s, unhealthy, very_unhealthy. The result shown that decision tree algorithm gives 91.9978 % accuracy which is more than that of Naïve Bayes algorithm viz. 86.663%.