Predict the Personality of Facebook Profiles Using Automatic Learning Techniques and BFI Test

Abstract

The present research work aims to predict the personality of a user’s Facebook profile. To do this, we have identified the attributes that are extracted from Facebook, with which the prediction of personality was possible. The data was extracted using the Graph API of Facebook, which was implemented in a web page. To achieve the knowledge base of machine learning, the BFI personality test is implemented for 118 users. In order to perform the training and classification of the automatic mode of learning by using the Weka tool, the degree of accuracy of the algorithms used in the prediction of the user’s personality was verified. The evaluation was carried out with two scenarios: using supervised learning and not using unsupervised learning. The work done yields results that indicate that it is necessary to increase the dictionary of data of the Spanish language, another result obtained is that in supervised learning, they gave data in which women have tendencies to be of neurotic personality compared to men. These data also determined that women are more difficult to predict their personality.