LAUSR

ABSTRACT In this paper, we analyse Markov prediction as a suitable model to suggest the next assembly step in the manufacturing process. The goal is a decision support system which can assist the workers in a factory, at least in their training period, to manually assemble a product. We evaluate the proposed context-based predictor on a dataset collected through an experiment involving 68 trainees and compare it with our previously implemented two-level context-based predictors. The experiment consisted of assembling a tablet composed of seven components. We were interested in finding the most efficient prediction method that can be soundly employed in real-life settings.