Research And Implementation Of Intelligent Information System For Catering Enterprises

With the rapid development of the Internet and artificial intelligence,data mining analysis methods have been continuously integrated into the research of various industries,and the catering industry has also entered the digital era.Facing the increasingly fierce competition in catering market,more and more catering companies implement digital management,thus accumulating a large amount of data.However,the information in the data has not been effectively used despite its important values.In order to meet the explicit needs of consumers,tap their potential needs,and guide their demand trends,in-depth analysis and mining of historical data are required to extract hidden information from various heterogeneous data sources.This paper conducts data analysis and forecast based on the internal sales data and external Internet public opinion data of catering companies,which could help to improve the management and operating level of catering companies.The main research work can be summarized as follows:1)This paper proposes a review text classification model based on Hybrid-Attention GRU neural network(denoted by HATT-GRU).It adopts character-based network structure to avoid grammatical restrictions.And by adding sentiment features to the model,the recurrent neural network can make full use of feature information in the text.Attention mechanism is applied to combine character attention and sentiment feature attention,so that the network can better focus on features that are more conducive to classification.Experiments show that HATT-GRU is superior to several current mainstream models on both Chinese and English review data.2)This paper presents a sales forecast model based on Self-Attention Dilated Convolutional Neural Network(denoted by SA-DCNN).The model can correctly handle chronological order and long-term dependencies without causing overly complex computation.At the same time,a multi-head self-attention mechanism is added to capture the correlation between the input time series and different features,and can learn multi-dimensional information in different representation subspaces.Experiments on multiple datasets prove that the SA-DCNN model can achieve a better improvement in sales forecast.3)This paper builds a visual intelligent information system for catering enterprises,which integrates multiple algorithm components to perform multi-dimensional analysis of catering data,and also provides a friendly interactive operation interface to guide related management and decision-making of catering enterprises.