TempRisk v5.0 optimizes accuracy and limits complexity by maximizing quality of model inputs (weather patterns). By optimizing method for determining which inputs matter and filtering those that do not, solution isolates signal from noise. Model inputs are based on statistical and meteorological filters applied to both historical and current weather data. Naïve Bayes framework coupled with Genetic Algorithm enables solution to predict probability of extreme temperatures.
This story is related to the following:Forecasting Software