Cost of Dietary Data Acquisition with Smart Group Catering

Modeling the cost in human-assisted data acquisition processes

“…Based on researches and investigations into real-world data acquisition scenarios [1][2][3][4] , the key components constituting general HADA processes are summarized and illustrated in Figure 1, including: data quality assessment, cost-quality relationship of human-assisted operations, process structure and manpower allocation strategy. This section will mathematically model each of these components.…”

“…The cost-quality relationship of human-assisted operations can be represented as a Cost-of-Quality (CoQ) relationship function, 𝑄 𝐸𝑀𝐻 , with its shape determined by the marginal cost characteristics of human-assisted operations. Based on the general understanding of marginal costs in economics [10] and our on-site observations during the field research of dietary data acquisition process [4] , we assume the existence of increasing marginal costs in the human-assisted operations of HADA scenarios: as the operation quality target rises, the marginal labor cost growth for achieving the quality target increases.…”

“…In previous study [4] , parameter values for the CoQ relationship functions in human-assisted operations were directly assigned based on on-site observations, making the CoQ functions static and incapable of reflecting the composition of labor costs. To address this issue, we introduce the Performance Shaping Factor (PSF) theory from the field of human reliability analysis [11][12][13] to decompose labor costs of human-assisted operations.…”

“…However, due to the complexity of data acquisition environments and limitations in technology, there exists a considerable variance in the degree of automation in data acquisition processes across different stages of digitalization. Extensive human assists are required in various scenarios such as complex image recognition, extraction of unstructured textural data, and recording of data in traditional production scenarios [1][2][3][4] . Compared to automated devices, the instability nature of manpower may lead to human errors and compromised data quality.…”

“…Compared to automated devices, the instability nature of manpower may lead to human errors and compromised data quality. Obtaining higher overall data quality necessitates additional human resource investments, resulting in substantial labor costs [4] . It can be argued that there exists a trade-off relationship between the quality of data acquisition and the associated labor costs.…”