OPTIMIZATION OF HYPER PARAMETERS IN MACHINE LEARNING TECHNIQUES FOR AIR QUALITY PREDICTIVE ANALYSIS

To reduce health related problems due to air pollution, there is a need of effective air quality prediction. In this regard, enhanced AQI (Air Quality Index) prediction machine learning models are proposed. Datasets from different domains like air pollution concentrations and meteorological data are collected and integrated. Machine Learning models such as k-Nearest Neighbors, XGBoost, Support Vector Machine and Decision Tree models have been effectively applied. Optimization of hyper parameters for various machine learning models has been carried out. From obtained results, it is observed that XGBoost gives better results compared to other models with least error rate of 1.6.