Modeling the cost in human-assisted data acquisition processes

Abstract

With the continuous advancement of digitization across various industries, the demand for data has significantly increased, making data production a critical component in different industry production chains. Throughout different stages of digitalization, there exists a considerable variance in the degree of automation in data acquisition. Human-assisted data acquisition (HADA) processes, such as complex image recognition, extraction of unstructured textual data, and recording of data in traditional production scenarios, often incur higher acquisition costs. The understanding of data acquisition costs holds great significance in making optimal resource allocation decisions to acquire high quality data for subsequent applications during industry digitalization processes. This paper, based on Cost of Quality (CoQ) and Performance Shaping Factor (PSF) theories, proposes a universal cost analysis framework model for general human-assisted data acquisition processes. We contend that there exists a trade-off relationship between data quality and labor cost in human-assisted data acquisition processes. Furthermore, we take dietary data acquisition as an example and conduct a quantitative analysis of the trade-off relationship between data quality and labor cost using the proposed framework.