A comparative study of prediction and classification models on NCDC weather data
Ibrahim Gad, Doreswamy Hosahalli
Abstract
Weather forecasting plays a significant role in different aspects of life such as in the operation of hydro-power plants, renewable energy, flood management, and agriculture. Recently, machine learning techniques have been used for weather forecasting for large periods of time, as it is more accurate than models based on physical principles. To address various problems, varieties of machine learning algorithms are applied in different fields. In our work, to examine whether these models are robust to predict National Climatic Data Center (NCDC) weather conditions, we carried out to compare newly emerging models with traditional meteorological models. In this paper, a set of the most common machine learning techniques are explored to generate robust weather forecasting model for long periods of time. Moreover, the combinations of all the model parameters are considered for simulations and the performance results of each method using a 10-fold cross-validation procedure are presented. The experimental results of the classifiers show that the decision tree CART, XGBoost and AdaBoost models exhibit better classification accuracy when compared with the other methods and for regression task, the linear regression method performs better in terms of R2 metric.