The Research And Application Of Prediction Model For Supermarket Passenger Flow

Nowadays,the rise of unmanned supermarkets has once again set off a wave of offline consumption,the effective forecast of unmanned supermarket passenger traffic can play a finishing role and promote the development of emerging industries.However,research in related fields is lacking.As an emerging industry,the unmanned supermarkets will have a unique shopping method and industry model that will directly lead to the difference in passenger traffic between unmanned supermarkets and other traditional offline supermarkets,thus forming a new type of data characteristics,therefore,for the forecast of passenger flow in the next week or one month of the unmanned supermarket,it is urgent to propose a new and applicable forecasting method.This paper focuses on the problem of accurate forecasting of passenger traffic in unmanned supermarkets,based on the characteristics of the statistics of passenger flow data that the amount of passenger traffic data is small and the complexity of volatility is strong,after research analysis and experimental comparison,a combination forecasting model of passenger flow in unmanned supermarket with strong constructability and outstanding applicability is proposed that combined the advantages of several machine learning algorithms and statistical analysis methods.The model uses cubic spline interpolation method to fill the missing values,and extends the original historical data for data fusion through correlation analysis,K-means clustering algorithm is used for clustering,and a cluster number selection method based on contour coefficients is proposed,the feature extraction stage is added to the discrete wavelet transform to split the waveform,and input it into the gradient boosting regression tree model for training,to strengthen the model during data processing and feature extraction.The research results show that the model not only has strong applicability to the prediction of unmanned supermarket passenger traffic,but also has an outstanding prediction effect,it solves the problem that the existing forecasting models for new and complex passenger flow data are not applicable and can not achieve accurate forecasting,at the same time,it has contributed a scarce power to the field of passenger flow forecasting of unmanned supermarkets.In terms of application,the decision-making tree model of unmanned supermarket marketing model is constructed based on the prediction of passenger flow,the model is based on CART algorithm and related attributes,and divide marketing mode into hierarchical structure output,the experimental results show that the model is usable.In engineering,based on the prediction of passenger flow and marketing model,the intelligent management system of unmanned supermarket is designed and implemented,it not only provides managers with a more timely distribution of passenger flow,but also better provides decision-makers with decision-making information that can be referenced.