Forecasting Winter Precipitation based on Weather Sensors Data in Apache Spark

Abstract

The proposed paper introduces an approach providing weather information on winter precipitation types using machine learning techniques. The proposed methodology takes as input the data received from weather sensors and in following the winter precipitation model aims at forecasting the weather type given three precipitation classes, namely rain, freezing rain, and snow, as registered in the Automated Surface Observing System (ASOS). To enable the proposed classification, six supervised machine learning models were selected: Naive Bayes, Decision Stump, Hoeffding Tree, HoeffdingOption Tree, HoeffdingAdaptive Tree, and OzaBag. Results depicted that all the models performed well in terms of accuracy and computation time, while some achieved even better outcomes. Specifically, among all six models, OzaBag presented the best classification results, followed by HoeffdingOption Tree.