Analysis of Personality Traits from Text Based Answers using HEXACO Model

Abstract

Analysis was made to propose an algorithmic method that is less complex and efficient in the prediction of personality for applying it to machine learning. Prior before applying the machine learning algorithm, Cronbach's Alpha is applied for testingthe questionnaire made usingHEXACO model to check the reliability of the factor and variables considered. By designing a method to change Cronbach's alpha, we aimed to analyse the influence of conflicting responses on the internal reliability of a dataset. Contrary to popular opinion, random reactions can inflate the alpha of Cronbach when their mean differs from that of the true reactions. Except in scales of both positive and negative polarity products, set answers inflate the alpha of Cronbach. There is not much effect on the effects of inconsistent answers by the amount of response groups. For the study, the mean score is calculated compared against the standard value using One sample test to identify there is a significant difference. The result indicates that there is no significant difference in mean score and standard value. It means that all the interviewees has reasonable level personality with respect to Honesty-Humility, Extraversion and Conscientiousness. For the factors; Emotionality, Agreeableness and Openness there is a significant difference in mean score and standard value. Through the Mean score calculated using One-Sample Statistics, it can be interpreted that the Interviewees have more than significant level of Agreeableness and Openness Personality but less Emotionality. This result is compared to the result of many HR professionals. To make the comparison of the resultPearson correlation method is applied, to know is there a significant relationship between the result given by HR managers and personality predicted using the HEXACO Model. The result indicates, there is a significant relationship between HR manager report and the HEXACO model algorithm constructed for personality prediction. Also, the estimated Pearson correlation value (0.819) indicates that there is 81.9% similarity in the result given by HR managers and the HEXACO model algorithm constructed for personality prediction.