Cost Analysis Of Dietary Data Acquisition Based On Smart Group Catering Systems

The requirement for dietary data management is becoming more and more urgent as the public commonly begin to take their health condition and nutritional intake seriously.Traditional methods of dietary data acquisition are poorly reliable and cost too much,which makes it difficult and incompetent to drive dietary management application.Smart group catering(SGC)systems has recently emerged and brought dawn and hope for this problem.Aiming at providing dietary management services for customers,canteens with Radio Frequency Identification(RFID)or Computer Vision(CV)based systems deployed have sprung up,claiming to achieve dietary data acquisition.Although RFID and CV based systems have almost realized automation in terms of dietary data acquisition,the data obtained are still influenced by many factors.On-site interview of canteens with SGC systems indicates that the average quality of data acquired is not satisfying.It does not only jeopardize the efficiency of the SGC system,but makes it harder to promote dietary-data-based health management services.This paper is set up upon our team's interview of a dozen of real canteens with SGC system deployed.Our goal is to discuss and analyze the current status of dietary data acquisition and concentrate on the root cause of dietary data's quality problem.This paper mainly contains three sections:For the first section,a real data set is organized based on all we can collect from our interviewee canteens,and is used throughout our study.Also based on the data set,we start from the operating procedure analysis of SGC systems and summarize the three challenges when using SGC systems for dietary data acquisition.After that we also use the data set to explore the mechanism of error data generation and data quality control during the acquisition process.All of them lead to the inevitability and rationality of extra investment when using SGC systems to acquire dietary data,and thus the necessity and urgency of further study on the investment are also proved.In the second section the way of modeling costs of dietary data acquisition with RFID and CV based SGC systems is presented.Started from extracting dietary information flow from system operating procedures,using lessons from ABC costing,key procedures of data acquisition in two systems are identified by eliminating factors which are irrelevant with information flow.Then costs are decomposed and categorized by whether it is related to staff operation.The model of staff operation equivalent manhour(EMH)and operation accuracy are constructed in RFID systems,while the model of sample number and recognition accuracy are also built in CV systems.After we define the model parameter value by human recognitive analysis of key procedures,we analyze the basic characteristics of the two systems.For the last part,a dish level simulation platform is constructed based on two SGC system models.Numeric analyses of the relationship between dietary data acquisition cost and factors such as target accuracy,canteen features are conducted in the order from general cost to major staffoperation-related cost(SORC),from static target accuracy to dynamic target accuracy,from baseline canteen model to general canteen model.Then case study is also presented based on four major canteen models from our data set.Finally,we add substantial discussion about model parameter sensitivity,efficient correction limits and dish standardization in order to generalize our study.Our study shows that human costs dominate in dietary data acquisition,and the underestimation of human costs is the major reason for data quality issues.Our study has strong positive reference value to exposing the current status of dietary data acquisition,exploring the method of lowering dietary data acquisition costs and enhancing data quality.