Development of statistical and machine learning models to predict the occurrence of radiation fog in Japan

This study develops statistical and machine learning models based on discriminant analysis, logistic regression a support vector (SVM) to predict the occurrence of radiation fog in Japan. The selection suitable set explanatory variables for was made using Akaike information criterion (AIC). accuracies three were measured compared. To determine optimum combination variables, temperature, humidity, wind speed, precipitation, sunshine, visibility data considered. Based root mean square error (RMSE) AIC values, best found include: presence speed during night, minimum temperature amount cooling humidity previous day, at 18:00. A comparison predictive selected variable showed that model produced critical success index (CSI, or threat score) 22.5, while SVM both better CSI results—with scores 33.8 38.2, respectively.