Application of neural networks for the prediction of lavender’s essential oil quality from environmental growing conditions and analytical data

The production of lavender has an important economic value in the cosmetic and pharmaceutical industries (Corbetta F. et al. 2001). Moreover, this crop has an ornamental importance for landscape. In order to improve production, we investigated the relationship among lavender’s essential oil quality, environmental conditions of the growing locations and analytical components of the essential oil. The aim of the study was the evaluation and classification of the effects of the environmental conditions (temperature, rainfall, insolation, soil traits) on oil quality and its prediction through neural networks (Richards et al. 1991), to be applied in the identification of the best areas for the cultivation of lavender. Another goal was the interpretation of the chemical compound pattern s of the oil samples for the prediction of the quality perceived by the panel test. By regression analysis the most important compounds for the perceived quality from sniffing analysis has been also identified.