The Cherry Stem Evapotranspiration Prediction Research Based On Agricultural Internet Of Things (IoT)

The explosion of global population,the diminishing arable lands and the increasing demands for foods have appealed for industrialization and intensification of agriculture.Meanwhile,the flourish of Internet of Things and affiliated data-driven techniques also offer new ideas for enhancing agricultural production and operations,where the prediction of stem sap flow is a paramount issue for crop production and irrigation systems.However,tremendous and unsynchronized data from agricultural cyber-physical systems bring large computational costs as well as make it impossible for performing conventional machine learning methods.To precisely estimate(i.e.predict)evapotranspiration with acceptable computational costs under the background of Io T,we combine time granulation computing techniques,Gradient Boosting Decision Tree and Bayesian Optimization(BO)to propose a hybrid machine learning approach.In the combination,a fuzzy granulation method and a time calibration technique are introduced to break voluminous and unsynchronized data into small-scale and synchronized granules with high representatives.Subsequently,GBDT is implemented to predict evapotranspiration and Bayesian Optimization is utilized to find the optimal hyper-parameter values from the reduced granules.Io T data from Xi’an Fruit Technology Promotion Center in Shaanxi Province,China verify that the proposed granular-GBDT-BO is effective for cherry’s evapotranspiration prediction with reduced computational time,acceptable and robust predictive accuracy.Besides,we also implement our proposed methodologies to perform agricultural Io T air temperature time series data prediction and scenario analysis,where the results demonstrate the generality of our proposed method.In terms of theoretical contributions,this manuscript proposes a general framework and a specific methodology for cherry tree stem Evapotranspiration prediction from data-driven perspective,which considers the characteristics of agricultural Internet of Things and complements stem Evapotranspiration prediction research stream under the background of agricultural Io T.Meanwhile,this general idea for handling Io T data can inspire industrial Io T data analytics,smart city,finance,and other domains which involves multi-sensors data analytics.When it comes to practical insights,the precise estimation of crop’s evapotranspiration could provide operational guidance for plant irrigation,plant conservations,and pest control in the agricultural greenhouse.