Analysis Of Factors Affecting Online Popularity Of Shanghai Catering Businesses

Mobile Internet technology,like an irresistible force,is infiltrating and expanding into all sectors of human society and providing unprecedented thrust to the world.In addition,the "food for the people" has penetrated into the national bone marrow,which means that China has a wider market relative to the rest of the world for the study of food and beverage.Therefore,in the face of the increasingly developed online consumer environment,it is particularly critical for catering businesses to make use of large data analysis techniques to achieve the transformation of the Internet Ecology.Therefore,the main work of this paper is to evaluate the index data of the gourmet merchants in Shanghai area,by selecting the total number of online consumer reviews as the dependent variables to measure the popularity of the catering merchants,the per capita price,taste,service,environmental scores,the number of stores,the star recommendation,the dish type,the region,and so on.Customs index is taken as an independent variable,and the page data of online catering business is selected as the data source.Second,the traditional multiple regression model was used to model the explanatory variables and the explained variables,and the significant analysis was carried out.The AIC criteria were used to screen the related variables.Third,the BMA model of the selected variables is constructed to measure the importance of each variable in the "real" model.Finally,the strong influence variables of the posterior probability of the BMA model are more than 20% as the input layer,and the BP neural network is used to train the variables,and the trainers are used to predict the popularity of the businessmen on the division line.Finally,through the analysis of the influencing factors and "priority" of the online catering consumption,the innovation scene,the consumer demand,the improvement of the restaurant model,the positioning of the restaurant,the help of social hot point,the production of communication content,to improve the competitiveness.To sum up,catering brands need to create content that has the power to transmit to target communities in attractive situations.This paper may make some contributions in the following aspects: by combining the traditional regression model with the BMA model,the defects of over-fitting of the traditional model and low efficiency of the BMA model are avoided;by measuring the importance of the "real" variables,the input variables of the BP neural network are screened,and the prediction accuracy is improved..