Abstract
Accurate prediction of environmental changes in a greenhouse is crucial for precise control and promoting crop growth. However, the microclimate environment in a greenhouse is nonlinear, temporal, multivariate, and strongly coupled, making it difficult to establish a robust fitting model. To address these issues, this study proposed a variable weight combination prediction model based on attention mechanism optimized Bi-directional Gated Recurrent Unit (BiGRU-Attention) and Light Gradient Boosting Machine (LightGBM). The Particle Swarm Optimization Algorithm (PSO) was employed to optimize the weight coefficients of the predicted values from BiGRU-Attention and LightGBM models at different times. This optimization aimed to enhance the accuracy of predicting air temperature, air humidity, and Photosynthetically Active Radiation (PAR) in a greenhouse. In predictions spanning time steps from 30 to 120 min, the variable weight combination prediction model demonstrated superior performance compared to single models and equal weight combination model BiGRU-Attention-LightGBM. At the time step of 120 min, the coefficient of determination R2 for air temperature was 0.9586, air humidity was 0.9232, and PAR was 0.8066. This indicated that the variable weight combination model (PSO-BiGRU-Attention-LightGBM) could more accurately predict the future dynamic trends of climatic environment factors in a greenhouse.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call