THE ENTROPY OF PREDICTING PERSONALITY FROM FUNNINESS SCORES

This investigation is a correlation and entropy study of two models of personality: The Big Five (B5) psychometric personality traits and the Age Trend Classification (ATC) model of artificial personality. ATC is based on a nonlinear dynamic model of animal motivation; the classification is computed a posteriori from funniness scores of arbitrary selected 100 humorous sentences. The purpose of the study is to assess the prediction of B5 traits by linear regression from funniness scores. The performance of prediction is measured by the Kullback-Leibler divergence (Kld) entropy of information loss. The analysis is implemented in stages, the first computes the correlation of participants mean funniness score with B5 traits. The mean is a measure of personal bias in assessing humour and includes contributions from multiple biases, such as mood, insecurity, personality attributes and emotional states. The analysis of the second stage is performed on scaled funniness scores and sentences with opposite orientation are aggregated and keyed, as is done in B5 assessment questionnaires. The graphic results show a repeating pattern of profiles not observed before. Stage three involves splitting the data among age groups, the graphs roughly showed a general trend of declining correlation with age, which suggests that the influence of motivation over traits in general also declines with age. The ATC model of artificial personality is intended as a counterpart of the personality of the user in a man-machine interface, where the machine could complement the user's personality in order to achieve higher learning, more entertainment and/or improve productivity