To effectively support individual contributors in large scale ideation settings, we need a computational understanding of their cognitive processes. Although related work exists that references models from psychology, there are two shortcomings in existing approaches: First, they only analyse ideas on a statistical level. Second, they lack a notion of individual ideator differences. This work proposes a new user model for ideation, based on process-models from psychology and knowledge-graph based information extraction from submitted idea-texts. By building a computational model of ideation, this work aims to enable new graph-based analysis methods of cognitive ideation processes and new approaches to adaptive ideation support systems. These systems could be used to detect and counter fixation in individuals and to guide group efforts based on categories that should be explored.