Construction And Application Of Cotton Yield Prediction Platform

Cotton is a pillar industry in Xinjiang,which plays an important role in economic and agricultural development.The formation cycle of cotton yield is long,the yield formation process is complex,and it is easily affected by environmental factors.Therefore,timely and accurate prediction of cotton yield is of great value and significance to Xinjiang’s cotton field management,agricultural decision-making,market control and other aspect.As the types and quantities of data used for cotton yield analysis are increasing,the traditional databases gradually show the problems of slow access speed,low efficiency,poor stability,and many restrictions on the types of data that can be stored when faced with these data.With the deepening of the research,more and more models have been used to predict the yield of cotton in different periods.When using these models to predict yield,users face the problems of difficult data transfer and complicated use process.Therefore,a comprehensive platform integrating cotton heterogeneous data classification storage and processing algorithm library management is built for users to achieve the purpose of integrating data resources,unifying data management,and realizing online data sharing.In this thesis,based on big data technology,a distributed database was built to store heterogeneous cotton data,which can better manage structured and unstructured cotton-related data in a unified manner.Secondly,a data processing algorithm library is built based on the Web Services system,and a web interface is built using the Vue.js framework to provide users with an entry to call the algorithm library,which simplifies the process for users to process cotton data.Finally,based on the Spring framework technology,a cotton yield prediction platform was built and developed,which realized the connection with the distributed database and algorithm library.Users can conveniently query and manage the data they uploaded themselves using this platform and can also easily call the cotton yield prediction model for rapid prediction of cotton yield.In this thesis,the performance of the distributed database is tested and compared.In the case of a single connection,the query performance of the distributed database is slightly lower than that of the centralized database,and the writing performance is basically the same,but the stability is stronger than that of the centralized database.When the number of concurrent connections gradually increases,The overall decrease in query and write performance of the distributed databases(1300 TPS and 1000 TPS)is lower than that of the centralized databases(4000 TPS and 3800 TPS);that is,the distributed databases constructed in this thesis can provide users with more stable services under the premise of ensuring performance.At the same time,this thesis also compares the prediction results of cotton yield prediction by invoking the model from the algorithm library with the prediction results obtained by using the code to run the model.The difference between the coefficient of determination(R~2)of the two methods is only0.0302%,and the difference between the root mean square error(RMSE)is only 0.0275 kg/mu.In other words,it can be considered that the prediction of cotton yield by calling the cotton yield prediction model built by the algorithm library in this thesis can achieve the effect of calling the same model for cotton yield prediction in the past.In this thesis,a distributed database is used to realize the classified storage of cotton heterogeneous data,which improves the data access performance,management efficiency,and reliability,and a data processing algorithm library is built based on the Web Services system,which provides users with a simple data processing flow and a simple calling method for the cotton yield prediction model.The framework technology is used to build a platform for cotton yield prediction that realizes the sharing and comprehensive utilization of data and the rapid prediction of cotton yield and has certain practical significance for Xinjiang cotton production and industrial development.