Open Access Open Access  Restricted Access Subscription Access

Forecasting Air Pollution Index in Relation to Stubble Burning in Punjab Using Machine Learning and Genetic Algorithm


Affiliations
1 AP, Khalsa College, Amritsar, India
2 Assistant Professor, Khalsa College, Amritsar, India
 

Tons of stubble is generated as residue after harvesting wheat and paddy agriculture crops. Farmers burn the stubble to dispose of it because there is not enough time for the next crop and take it as a quick and inexpensive solution. It harms our ecology and eco-system in numerous ways. For the survival of humanity, it is crucial right now in India to track and forecast the Air Quality Index (AQI). The most major and dangerous air contaminant in this area is particulate matter (PM). For making predictions, machine learning (ML) technology is more effective than earlier conventional methods. Numerous ML algorithms, such as Random Forest (RF), Support Vector Machine (SVM), classification methods, Regression analysis, etc., were widely used for maximum prediction. But here by using Random Forest with Genetic Algorithm (GA) a hybrid approach prediction takes place much better. In order to improve the output of the data-adaptive computation, GA was used. Presently, data on air pollution from the preceding five years have been analyzed and forecasted for a study on Punjab's key cities to estimate and forecast PM concentrations. It was examined how PM concentrations vary with the seasons and some air pollutants. Variable importance ranking (VIR) was used to assess the effectiveness of the presented model. Here, the main emphasis was on taking into account some of the data sets from important cities in Punjab for the prediction of ambient pollution and air quality by using machine learning with genetic algorithm. Various common metrics were used to compare the results of all the strategies.

Keywords

Particulate Matter (PM), Air Quality Index (AQI), Correlation Analysis, Machine Learning (ML), Variable Importance Ranking (VIR), Random Forests (RF), Support Vector Regression (SVR).
User
Notifications
Font Size


  • Forecasting Air Pollution Index in Relation to Stubble Burning in Punjab Using Machine Learning and Genetic Algorithm

Abstract Views: 310  |  PDF Views: 0

Authors

Dr. Rachhpal Singh
AP, Khalsa College, Amritsar, India
Dr. Rupinder Singh
Assistant Professor, Khalsa College, Amritsar, India
Prabhjot Kaur
Assistant Professor, Khalsa College, Amritsar, India

Abstract


Tons of stubble is generated as residue after harvesting wheat and paddy agriculture crops. Farmers burn the stubble to dispose of it because there is not enough time for the next crop and take it as a quick and inexpensive solution. It harms our ecology and eco-system in numerous ways. For the survival of humanity, it is crucial right now in India to track and forecast the Air Quality Index (AQI). The most major and dangerous air contaminant in this area is particulate matter (PM). For making predictions, machine learning (ML) technology is more effective than earlier conventional methods. Numerous ML algorithms, such as Random Forest (RF), Support Vector Machine (SVM), classification methods, Regression analysis, etc., were widely used for maximum prediction. But here by using Random Forest with Genetic Algorithm (GA) a hybrid approach prediction takes place much better. In order to improve the output of the data-adaptive computation, GA was used. Presently, data on air pollution from the preceding five years have been analyzed and forecasted for a study on Punjab's key cities to estimate and forecast PM concentrations. It was examined how PM concentrations vary with the seasons and some air pollutants. Variable importance ranking (VIR) was used to assess the effectiveness of the presented model. Here, the main emphasis was on taking into account some of the data sets from important cities in Punjab for the prediction of ambient pollution and air quality by using machine learning with genetic algorithm. Various common metrics were used to compare the results of all the strategies.

Keywords


Particulate Matter (PM), Air Quality Index (AQI), Correlation Analysis, Machine Learning (ML), Variable Importance Ranking (VIR), Random Forests (RF), Support Vector Regression (SVR).

References