Forecasting Daily Urticaceae Pollen Count By Artificial Neural Networks

Pollen and spore forecasting has become an important aim in aerobiology and is one of the most studied topics. Air pollen forecasting is directly related to a regional agriculture, environment, health and other aspects of people’s lives. The most used tools for this problem are regression models. The advanced mathematical methods can be applied to the problems that cannot be solved in any other effective way, and are suited to predicting the concentration of Urticacea airborne pollen in relation to weather conditions. A few works have used more sophisticated methods such as Artificial Neural Networks (ANNs). In this paper, we developed some of ANNs models to forecast Urticaceae pollen concentrations for pre-peak, post-peak and whole periods in the atmosphere of Tirana, Albania.

This method gave good results for Pearson’s correlation and R-square: the correlation was 0.98, R-square was 0.82 for the pre - peak period and the correlation 0.97, R-squared 0.85 for whole correlation, respectively. We used different MPL with three layers and a various number of neurons in the hidden layer. Experimental results show an advantage of the ANNs against statistical methods, although there is still room for improvement. Used models gave more satisfactory predictive results, where it was best for the pre-peak, then for the whole period and weak for the post-peak period.