CLASSIFICATION OF PEOPLE BY PSYCHOLOGICAL PERSONALITY TYPES BASED ON THE HISTORY OF CORRESPONDENCE
Ключевые слова:
social networks, ML, NLP, psychotype recognition, Keirsi temperament model, classification.Аннотация
Temperament is a set of innate tendencies of the mind associated with the processes of perception,
analysis and decision-making. The purpose of this article is to predict the psychotype of individuals based on chat
stories and follow the Keirsi model, according to which the psycho type is classified as a craftsman, guardian,
idealist and mind. The proposed methodology uses a version of LIWC, a dictionary of words, to analyze the context
of words and uses supervised learning using KNN, SVM, and Random Forest algorithms to train the classifier. The
average accuracy obtained was 88.37% for artisan temperament, 86.92% for caregivers, 55.61% for idealists, and
69.09% for rationality. When using the binary classifier, the average accuracy was 90.93% for artisan temperament,
88.98% for caregivers, 51.98% for idealism, and 71.42% for rationality.
