Temperature and relative humidity prediction in South China greenhouse based on machine learning

Prediction of the greenhouse temperature and relative humidity is very important, which can forecast the environment parameters for manual intervention in advance. However, temperature and relative humidity prediction systems face two critical limitations: inconsistent temporal resolution in data acquisition and the absence of standardized protocols for environmental data collection, which collectively lead to non-uniform control strategies that compromise system interoperability in agricultural applications. This research predicted the temperature and relative humidity with different time interval in South China greenhouse by the model of BPPSO, LSSVM and RBF, which has proved their superiority in temperature and relative humidity prediction. The results showed that the R² of temperature and relative humidity increase gradually with the decrease of the time interval, and the time interval of 15 min got the maximum value. The R² of the temperature predicted by three models were 0.923, 0.923,0.912, and the R² of the relative humidity were 0.948,0.952, and 0.948, respectively. The prediction accuracy of relative humidity was higher than that of temperature. All three models could be used to predict temperature and relative humidity in greenhouses in South China, among which LSSVM had higher R² than the other two models. When the time interval was 15 min, the MAE, MAPE and RMSE of temperature were 0.574, 1.941 and 0.867, respectively, while the relative humidity of that were 2.747, 3.383 and 3.907, respectively. It concluded that the LSSVM model with time interval of 15 min was suitable to predict the temperature and relative humidity in south China greenhouse. This study provides reference for early intervention of greenhouse temperature and relative humidity management.