Social network profiling for cultural heritage: combining data from direct and indirect approaches

The work argues for quick profiling methods from social networks for use in cultural heritage applications. Explicit (inquiries about user actions, like game playing) and implicit (observations from user actions on social networks) methods are tested, in an attempt to extract user personality profiles and in particular cognitive style profiles, using the MBTI tool. Qualitative and quantitative approaches have been applied to validate the results. So far, it seems that users’ cognitive profiles can be predicted from social media observations and user actions (i.e., playing games) for 3 out of the 4 MBTI dimensions. There seem to be relatively accurate predictions for the dimensions Judging–Perceiving and Extraversion–Introversion. Sensing–Intuition is a little more difficult to predict. Currently, the Thinking–Feeling dimension cannot be predicted from the existing data. Future works will concentrate on improving the prediction rate for the Sensing–Intuition dimensions and discovering ways to predict the Thinking–Sensing dimension from social network information.