Growth Model of Greenhouse Cucumber Based on Random Forest Bayesian Optimization
To solve the low production level of greenhouse cucumber,difficulty in precise regulation of greenhouse environment,and insufficient intelligent application,this study is based on environmental data and cucumber growth data in the greenhouse of Shandong Agricultural University Science and Technology Innovation Park Horticulture Experiment Station.Random Forest Bayesian Optimization Algorithm(RF-BO)is used to construct growth simulation models for greenhouse cucumber during the planting period,vine extension period,initial flowering period,and harvesting period,and compared with the growth model established by Random Forest(RF)algorithm.The results show that the simulation effect of the greenhouse cucumber growth model based on RF-BO is better than that of the greenhouse cucumber growth model constructed by RF algorithm in each development stage.The determination coefficient R2 of the growth models in each development stage is above 0.9.The root error RMSE range is between 0.121 and 0.317,and the average absolute error MAE range is between 0.096 and 0.221,which can accurately simulate the growth dynamics of greenhouse cucumbers.The greenhouse cucumber growth model constructed by this research institute also provides reference for the growth prediction of other horticultural vegetables or crops,thereby providing more scientific and feasible decision-making basis for precise environmental regulation and agricultural production.
国家科技期刊平台
NSTL
万方数据
