Research On Smart Agriculture Decision-making System Based On Machine Learning Algorithm

With the continuous development of the Internet of Things,artificial intelligence and big data technology,smart agriculture has become a hot topic in current agricultural science and technology research.Smart agriculture is supported by the Internet of Things technology.It integrates information,sensing and wireless communication technologies.It relies on various sensing nodes and wireless sensor networks deployed in agricultural production sites to complete the comprehensive sensing and reliable transmission of agricultural information.A new model of agricultural development that is intelligently handled.It not only improved the way of agricultural production management,improved the efficiency of agricultural production,but also promoted the sustainable and stable development of agriculture.However,the problem of the low precision of automatic control in intelligent agriculture and the difficulty of information decision-making risk within the scope of the requirements is a problem that needs to be solved.For this reason,this thesis has carried out the following research work:Firstly,a threshold optimization method based on machine learning is proposed.The method uses the Bayesian statistical inference mechanism to learn the data of agricultural production environment,establishes the multidimensional Gaussian statistical model under each environmental parameter index,and uses the EM algorithm to establish the joint distribution posterior probability density likelihood function to obtain the sparse item + regular term.Log-likelihood function model;environmental parameter optimization decision threshold obtained by ADMM optimization and NP decision rule.Secondly,a software platform for intelligent decision-making system based on machine learning was designed and developed.This platform uses key technologies such as Qt and cloud computing to implement the management center module,data center module,map center module,and equipment center module of the smart agricultural decision-making system.main content.The system can complete the user's login and registration,real-time monitoring and display of agricultural production environment parameters,changes in the corresponding parameters of the trend,query and analysis of historical data,map display and greenhouse environmental parameter status information monitoring and other functions.Finally,the proposed machine learning-based threshold optimization method is applied to the decision-making system of intelligent agriculture.Through the threshold optimization method based on machine learning,combined with the historical data in the agricultural production site,the optimal decision threshold for each environmental parameter is obtained.The equipment control center,which guides the smart agricultural decision-making system,can remotely and adaptively drive the corresponding electrical equipment enable.This research can achieve precise and intelligent control while reducing the risk of decision making.The research results were applied in the smart agricultural greenhouse in Ili,Xinjiang,which increased productivity by 43% and created an additional 32% of economic value.This result can also be applied to other operational smart agricultural systems.