Applying machine learning to forecast daily Ambrosia pollen using environmental and NEXRAD parameters.

摘要: Approximately 50 million Americans have allergic diseases. Airborne plant pollen is a significant trigger for several of these Ambrosia (ragweed) known its abundant production and potent effect in North America. Hence, estimating predicting the daily atmospheric concentration (ragweed particular) useful both people with allergies health professionals who care them. In this study, we show that suite variables including meteorological land surface parameters, as well next-generation radar (NEXRAD) measurements together machine learning can be used to estimate successfully concentration. The supervised approaches included random forests, neural networks, support vector machines. performance training independently validated using 10% data partitioned holdout cross-validation method from original dataset. forests (R= 0.61, R2= 0.37), machines (R= 0.51, R2= 0.26), networks (R= 0.46, R2= 0.21) effectively predicted pollen, where correlation coefficient (R) R-squared (R2) values are given brackets. Three independent approaches—the coefficients, interaction information—were employed rank relative importance available predictors.