Efficient weighted naive bayes classifiers to predict air quality index

In the past few decades, there have been many environmental changes that have lead to deteriorating air quality influenced by number of criteria pollutants and the prevailing climatic conditions. These pollutants lead to respiratory problems and other environmental effects such as acid rain, greenhouse effect etc. Therefore, quality prediction of air quality has long served as prospective and practical study area that has received massive attention. In literature, Naive Bayes classifier based on independence and equal importance assumption has been used by many researchers for air quality prediction. However, these assumptions never hold practically. To fulfill this purpose, two classifiers based on Weighted Naive Bayes named as Covariance based Weighted Naive Bayes and Convergent Cross Mapping based Weighted Naive Bayes have been proposed in this study to predict the air quality index of Faridabad, Delhi and Gurugram in India. The findings of the experimental work conducted in this study showed that both the proposed weighted classifiers perform better than the traditional Naive Bayes, Support Vector Machines and Neural Network classifiers with respect to various performance metrics- accuracy, average precision, average recall, error rate and F1 score. Further, this study depicts that Covariance based Weighted Naive Bayes and Convergent Cross Mapping based Weighted Naive Bayes have an average accuracy of 83.6% and 82.12% respectively.