Predicting Air Pollution Level in Particular Area Using KNN by Comparing Accuracy with SVM

Abstract

To predict the air pollution level in a particular region area using a K-Nearest Neighbor algorithm compared with the Support Vector Machine algorithm, The Novel K-Nearest Neighbor Algorithm and the Support Vector Machine Algorithm are two groupings. The algorithms were implemented and evaluated on a dataset of 32516 records. Various air pollution was identified through a programming experiment with N = 5 iterations for each method. G power is set at 80%. The confidence interval is 95%, and the threshold value is 0.05%. The G-power test is around 80% accurate. The K-Nearest Neighbor algorithm (97.44%) has better accuracy when compared with the Support Vector Machine (70.32%). The K-Nearest Neighbor algorithm has the highest accuracy compared to the Support Vector Machine algorithm. In the prediction of air pollution, K-Nearest Neighbor has better performance when compared to the Support Vector Machine Algorithm.