MACHINE LEARNING TECHNIQUES FOR AIR QUALITY PREDICTION

Air pollution poses a major threat to the society. Monitoring the quality of air is the key to assure public health. Naïve and empirical methods for predicting air pollution is not accurate. This work proposes to develop an air pollution prediction system that can predict the quality of air using machine learning techniques. A review is done to study the different machine learning approaches conducted by researchers. An architecture is developed which takes in input data and follows different stages like pre-processing, learning and evaluation. A model can be developed using the training data. The model is then tested using the test data. Air pollution can then be predicted on the test data to know how accurately the model predicts. Various techniques like Logistic Regression, Random Forest, Support Vector Machine and Neural Networks are discussed for constructing the model to be used for prediction.