Machine learning techniques for weather forecasting
Sanders, William Samuel
MetadataShow full item record
Machine learning models were developed in order to forecast weather variables such as solar radiation, temperature, and wind speed for one to 24 hours in advance. Weather predictions and ground truth weather observations were sourced from the National Oceanic and Atmospheric Administration (NOAA) and the Georgia Automated Environmental Monitoring Network (GAEMN) for five cities in Georgia. Results indicate that incorporating weather forecasts becomes increasingly more important for accurate solar radiation prediction at longer prediction windows, and also that postprocessing of NOAA's weather forecasts can drastically improve accuracy beyond usage of the raw forecasts alone.