Study On Prediction Of Catering Industry Food Sales

With the development of modern information technology and scientists continuously optimized mathematical model,Information management is gradually applied to the food and beverage industry,The traditional catering industry management model is gradually developed with information technology and automation management.Food and beverage industry for many years of accumulated data has not been fully utilized,resulting in a waste of data,but provides data sources to current study.Modern management catering industry has matured,But due to the large backlog of dishes,which cause rotting waste,thereby reducing profits or insufficient supply of dishes,these can’t meet consumer demand and result in decreased turnover,which cause customer satisfaction declined.Sales of daily dishes restaurant will be a more accurate prediction to solve the above problems,thereby increasing the competitiveness of the restaurant.I passed on the automated management system used in the domestic restaurant industry study and found that Catering management system only related to a simple inventory management,and they not related to the sales forecast dishes,which can not predict sales based on existing information resources.Since the dishes are complex sales trends of model uncertainty,which associated with the establishment of predictive models is not easy,most of the applications forecasting model currently used on some of the more long-term observation activities,even if used on short-term forecasting activities,each specific the model has a great difference.I was inspired by other industry-related prediction models,respectively Based on factors of division and classification,establishment and optimization algorithm prediction model to test the feasibility of the model and to improve prediction accuracy.Neural network have good nonlinear fitting ability [1],which has been widely used in various fields.More and more training and comprehensive forecast data forecast based on neural network.In the food and beverage industry,we can get a lot of valid data resources through previous sales records and related data.However,since the factors are large and complex,and there is a definite relation between these factors,it is not satisfied for some more information and comprehensive data,which can’t be used to predict the neural network.The popular in other industries prediction algorithm is BP neural network,compared with the BP neural network,ELM running speed has a great advantage,and the error is relatively smaller.At present,the problem of insufficient data,a better approach is gray system.Food sales in the past dishes Classification rough aging prediction accuracy is not enough,a single prediction methods,the best prediction error are ± 20%,Companies can’t reach the desired results,this paper based on experimental prediction data predicted the relationship between the data for the main trail,through the experimental data by a more detailed treatment,study each factor level and the level of influence factors effect,By BP neural network,ELM and the gray system prediction dishes on the restaurant industry sales,the purpose is to complete all sorts of dishes catering business short-term sales forecast.Through a variety of dishes sales forecasting model was established and contrast,it aims to establish high precision prediction model for prediction of dishes catering industry,food and beverage companies can achieve short-term sales forecast all sorts of dishes,and minimize the prediction error,which really apply and implement to the catering industry,reducing inventory,saving operating costs,thereby increasing the competitiveness of enterprises.