Research On Food Sales Forecast In Catering Industry Based On Deep Learning

With the substantial improvement of China's economic strength,the improvement of people's living standards has spurred people to pursue a higher level of material life.The injection of technological elements has promoted the intelligent transformation of the catering industry.The increasing number of restaurants in catering enterprises and the high closing rate make the competition in catering industry increasingly fierce,so the development of enterprises should find new ways according to the current and future industry trends to find a suitable business model for catering industry at the present stage.In addition,food procurement is closely related to corporate profits for catering companies.Therefore,building a high-accuracy food sales forecasting model is of great significance to the good development of the company.This thesis analyzes the food sales data of a well-known catering company for two years and seven months,and finds the different characteristics of the food sales during normal working days and holidays.The sales on ordinary working days follows a sinusoidal distribution,and the period is seven working days.The sales during the holidays are generally higher.Based on this,this thesis builds a sales forecasting model using deep learning for catering industry.The model is divided into an ordinary working day sales forecasting model O-Model and a holiday sales forecasting model H-Model.In this thesis,we use two years of data to carry out model training and adjustment,and the last 7 months of sales data are used for model evaluation.The experimental results show that compared with the traditional time series.model autoregressive moving average model(ARMA)and machine learning algorithm Xgboost,the use of deep learning for food sales forecasting has more applicability and can obtain more accurate prediction results.The hardware cost of this model is not high,and there is no need to worry about real-time performance.So that it can be effectively applied to the food sales forecast.Finally,this thesis designs and implements the catering industry food sales forecasting system,which is convenient for viewing the sales forecast and historical sales of the food,and provide a scientific basis for the company to purchase food and make decisions,thereby strengthening the competitiveness of the catering enterprise and promoting the sustainable development of the enterprise.