The health of millions of people around the world is impacted by air pollution, a serious environmental issue. PM2.5, a type of fine particulate matter found in the air, is a significant source of air pollution and has been associated with a number of health issues.Sources includes industrial activities, transportation,fuel uses, agriculture etc. Common pollutants include Particulate Matter(PM),Nitrogen Dioxide(NO2), Sulphur Dioxide(SO2),Carbon Monoxide (CO).Particulate matter contains microscopic solids or liquid droplets that are so small that they can be inhaled and cause serious health problems. Of these, particles less than 2.5 micrometres in diameter, also known as fine particles or PM2.5, pose the greatest risk to health. Machine learning methods have recently demonstrated considerable potential in predicting PM2.5 levels. In this study, stacked and single-based supervised machine learning methods are suggested for PM2.5 prediction. Utilising actual data on air quality, the proposed methods are assessed. The findings reveal that, in terms of predictive accuracy, the stacked approach outperforms the single-based approach.It compromise of four different machine learning models respectively linear regression, support vector machine, artificial neural network and dicission tree, combined such order to achieve best possible accuracy till date.
SUBMITTED BY :
1.AKANSHA DAS(190310007006) SHRABANI MEDHI
2.JITJYOTI SHARMA(190310007021) ASSISTANT PROFESSOR
3.PALLAV KASHYAP(190310007033) CSE DEPT,GIMT,AZARA
GUIDED BY :