An intelligent fuzzy system to manage western flower thrips population under biological treatment in roses yield.

Experimentation centers in the French Riviera (SCRADH, Florisud, and others) are keen on informing growers about the propitious weather conditions relative to pest risk growth. The only existing method used nowadays is sending notifications to their mobile phones. In this study, we are interested in developing a system for Western Flower Thrips (WFT) risk assessment. We propose a fuzzy logic approach to build an intelligent system that supplies rose growers with an estimate of the next day's risk in their greenhouses. The system relies on today's data (data of the day) to predict the next day's risk, under the presence of bio-control agents, in particular, harmless fungi called Beauveria Bassiana. We could recapitulate the study's outcomes in the following elements. The most significant aspect is estimating a daily risk index that provokes premature detection of WFT. Besides, considering the utilization of biological treatment aids in reducing the pulverization of pesticides. In addition, the system incorporates a small number of sensors usually found in almost all greenhouses. Thanks to continuous monitoring of the weather conditions (real-time system), the system could help mitigate the percentage of yield loss. Finally, this is the first attempt to employ fuzzy logic to predict WFT subjected to biological treatment. We compared the obtained results to actual data. The model's accuracy was determined by three statistical indicators (coefficient of determination (R2), root mean squared error (RMSE), and mean absolute error (MAE)) showing a noticeable performance. The fuzzy model showed satisfactory results (R2=0.9, RMSE= 10%, and MAE=7.3%) making it reliable and robust for pest monitoring.