The advanced statistical methods in aerobiological studies

Agnieszka Grinn-Gofroń,Beata Bosiacka

doi:10.5586/aa.2012.023

Abstract

Pollen and spore forecasting has become an important aim in aerobiology. The main goal is to provide accurate information on biological particles in the air to sensitive users in order to help them optimize their treatment process. <br />Many statistical methods of data analysis are based on the assumptions of linearity and normality that often cannot be fulfilled. The advanced statistical methods can be applied to the problems that cannot be solved in any other effective way, and are suited to predicting the concentration of airborne pollen or spores in relation to weather conditions. The purpose of the study was to review some advanced statistical methods that can be used in aerobiological studies.

Highlights

Modern epidemiological studies from various countries indicate that currently 15–20% of the average population suffers from allergic diseases
The small sizes of these allergens allow for deep penetration of the bronchial tree, which often leads to allergic reactions of the lower respiratory tract
It is estimated that sensitization to pollen and fungal allergens relates to the growing number of people

Summary

INTRODUCTION

Modern epidemiological studies from various countries indicate that currently 15–20% of the average population suffers from allergic diseases. Pollen grains and fungal spores are one of the main sources of inhalation allergens. It is estimated that sensitization to pollen and fungal allergens relates to the growing number of people. This percentage is higher in the child population compared with the adult population (Burge , 2002; Beggs , 2004). The important direction of aerobiological studies is to seek correlations between the characteristics of the pollen or spore season and weather variables. Due to the complexity of the study object (a large number of analyzed variables, very irregular changes in the concentration of airborne pollen or fungal spores of a large variety of species, nonlinear correlations between parameters), multi-dimensional techniques and other advanced statistical methods of exploring data are preferred

Currently developed forecasting models

Findings

Multivariate regression trees