Personality trait prediction by machine learning using physiological data and driving behavior

This article explores the influence of personality on physiological data while driving in reaction to near crashes and risky situations using Machine Learning (ML). The objective is improve assistance systems considering drivers’ characteristics. Physiological behavioral were recorded sixty-three healthy volunteers during urban analyzed 5 ML algorithms discriminate driver’s according Big Five Inventory STAI trait. Seven step process was performed including pre-processing, Electrodermal Activity (EDA) time windows selection (one by one backward forward approach comparison with a pseudo-wrapped), traits assessment, input parameters optimization, algorithm trait cluster prediction. ROC Area Under Curve (AUC) used describe improvement. pseudo-wrapped/all possibilities method resulted 8.3% average for all (% AUC approach). detection ranged between 0.968 0.974 better Openness, Agreeability Neuroticism. Use association Neuroticism, Extraversion Conscientiousness previously defined literature slightly (maximum 0.961 0.993 cluster). Results are discussed terms contribution aids. study first use machine learning techniques detect measures context. Additionally, it questions optimization approach, selection, as well clustering • Considers different contexts uses data. Assesses relative types bio-signals accuracy. Highlights capacity classify driver traits. Tests benefit associations.