Research And Application Of The Temporal Feature-correlation Model For Sugar Raw Materials Demands Prediction

Sugar industry is the basic industry of China's food industry which plays an indispensable role in China's national economy.However,most domestic sugar companies suffer from low production efficiency due to the imbalance between supply and demand of raw materials.Through the accurate prediction of the demand for sugar raw materials,not only can we make more scientific and reasonable raw material procurement plans for sugar companies to ensure production balance,but also provide data services for raw material demand forecasting systems,and effectively improve the automation level of sugar enterprise management.However,most of the existing demand forecasting studies use traditional regression,linear programming and other algorithm models,and have not considered the characteristics of sugar raw materials business,ignoring the influence of time factors on raw material demand,resulting in limited accuracy of demand forecasting.Therefore,how to scientifically and accurately predict the demand for sugar raw materials is one of the fundamental and key issues that must be solved for sugar enterprises to realize modernized intelligent management.This paper studies the accurate prediction of raw material demand based on the supply and demands of raw materials in a real sugar company.In this paper,the Elman neural network are optimized through Modified Cuckoo Search(MCS)algorithm and the temporal features.The Temporal Feature-correlated MCS-Elman Neural Network(TMCS-ENN)is proposed.Firstly,the basic theory including time series,Elman neural network and cuckoo search algorithm are described,and analyzes the time characteristic correlation factors affecting the demand of raw materials in combination with the actual situation of sugar enterprises.Simultaneous sequence prediction algorithm is used for comparative analysis to determine the prediction model of sugar raw material demand.Secondly,based on the established MCS improved Elman neural network prediction model,the holiday characteristics associated with time characteristics are added to make the prediction model more compatible with sugar production.The business characteristics of the enterprise are analyzed and verified experimentally.Finally,for the real enterprise needs,a sugar raw material demand forecasting system based on the TMCS-ENN model is designed and implemented.The experiment proves that the model of this paper effectively improves the accuracy of the demand forecast of sugar raw materials in enterprises,and significantly improves the work efficiency and automation level of the procurement business of sugar enterprises.