Comprehensive prediction of multiple low temperature and high humidity diseases in greenhouse cucumbers based on...

Facility agricultural greenhouses are prone to forming low-temperature and high-humidity environments, which frequently induce the concurrent outbreak and spread of fungal diseases such as downy mildew, gray mold, and black spot disease. However, existing predictive studies often focus on individual diseases or environmental factors, making it difficult to meet the precise disease control needs in production settings with complex diseases. To address this issue, this study constructs an Infection Suitability Index (ISI) that quantifies humidity suitability using Beta functions and piecewise linear functions, enabling nonlinear coupling of temperature and humidity. With this as the core input, multiple linear regression (MLR), support vector regression (SVR), and MLR-SVR hybrid predictive models are developed. The results indicate that ISI is significantly negatively correlated with the incubation period of all three diseases. The MLR-SVR hybrid model demonstrated the best prediction performance, with a coefficient of determination (R²) ranging from 0.962 to 0.982 and prediction accuracy between 88.5% and 92.3%. It shows excellent applicability in greenhouse environments. This study provides a new method for multi-disease collaborative early warning in facility cucumber cultivation, and the proposed ISI index along with the hybrid model system has important practical value for intelligent pest control and precision agriculture.