Investigating Pollen Data with the Aid of Fuzzy Methods

The analysis of pollen data coming from a large number of monitoring sites and related to various pollen types is a multivariate knowledge discovery problem. The relationships between the variables and the forecasting of their behaviour may successfully be performed with the aid of computational intelligence methods. The present paper deals with data coming from 25 monitoring sites in Austria, including time series of pollen counts for 10 pollen types. Fuzzy rules and fuzzy clustering were employed, as well as appropriate graphical representation, for the analysis of the behaviour of the pollen types. Results indicate the ability to extract knowledge and to forecast pollen count levels that directly affect the quality of life of allergic people.