This study proposes a labor hour estimation model that does not rely on measured data, aiming to mitigate reliance on individual expertise and promote skill transfer within small and medium-sized manufacturing enterprises (SMEs). Leveraging the SECI model framework, the research develops a predictive model that simulates techniques used by skilled workers. This is achieved by integrating explicit representations of tacit knowledge with historical estimation data into a polynomial regression model. The proposed model demonstrated high predictive performance, achieving an accuracy rate of 75.26% within a ±10% error margin in evaluation experiments. These results indicate the feasibility of constructing an objective and reproducible estimation process even in environments where real-time measurement data is limited. In conclusion, this study contributes to the standardization and transparency of labor hour estimation practices and supports the development of a sustainable framework for knowledge transfer.
KIMURA et al. (Wed,) studied this question.