Stepwise selection of functional covariates in forecasting peak levels of olive pollen

摘要: High levels of airborne olive pollen represent a problem for large proportion the population because many allergies it causes. Many attempts have been made to forecast concentration pollen, using methods such as time series, linear regression, neural networks, combination fuzzy systems and functional models. This paper presents logistic regression model used study relationship between different climatic factors, on this basis predict probability high (and possibly extreme) selecting best subset variables by means stepwise method based conditional likelihood ratio test.