Abstract
Construction labour productivity is a cost efficiency measure of crews in producing outputs usually at an activity level. The relationship between the factors affecting the efficiency of crews and the achieved labour productivity is being studied using various stand-alone modeling approaches like regression analysis, neural networks, and fuzzy logic-based expert systems. However, the developed models suit only a specific context and most importantly, a method for transferring and generalizing the knowledge captured in the various models has not yet been fully developed. This paper presents the application of a granular fuzzy modeling approach for transferring captured knowledge and the process of developing a granular generalized construction labour productivity model having an improved prediction capability. The granular fuzzy model abstracts three construction productivity models dealing with industrial welding activities using a case-based reasoning approach on clusters of the respective model input data prototypes. The performance of the model is evaluated using coverage and specificity plots and different model parameters are optimized