Semantic networks have been used extensively in psychologyto describe how humans organize facts and knowledge inmemory. Numerous methods have been proposed to constructsemantic networks using data from memory retrieval tasks,such as the semantic fluency task (listing items in a category).However these methods typically generate group-levelnetworks, and sometimes require a very large amount ofparticipant data. We present a novel computational methodfor estimating an individual’s semantic network usingsemantic fluency data that requires very little data. Weestablish its efficacy by examining the semantic relatedness ofassociations estimated by the model.