A Comparison of Personality Prediction Classifiers for Personnel Selection in Organizations Based on Industry 4.0

A Comparison of Personality Prediction Classifiers for Personnel Selection in Organizations Based on Industry 4.0

Roberto Contreras-Masse, Juan Carlos Bonilla, Jose M. Mejia, Alberto Ochoa
DOI: 10.4018/978-1-7998-4730-4.ch012
(Individual Chapters)
No Current Special Offers


Nowadays, the internet has an astonishing amount of useful material for personality mining; nevertheless, many companies fail to exploit the information and screen job candidates using personality tests that fail to grasp the very information they are trying to gather. This research aims to highlight and compare the different machine learning classifiers that can be used to predict the personality of a Spanish-speaking job applicant based on the written content posted on their social networks. The authors conduct experiments considering the most critical measures (such as accuracy, precision, and recall) to evaluate the classification performance. The results show that the random-forest classifier outperforms the other classifiers. It is of utmost importance to correctly assess the resumes to determine the most qualified people in a smart manufacturing position.
Chapter Preview

Machine Learning and e-recruitment systems are booming nowadays (De Meo, Quattrone, Terracina, & Ursino, 2007); however, many of them don’t focus on user personality despite the huge amount of user data available. This appears as a good opportunity to broaden the existent recruiting systems’ boundaries; also, such a tool could be of value to recruitment centers that can’t afford to screen all the candidates’ personalities through personal interviews. Faliagka et al. (2012) created an e-recruitment system that included automatic personality mining for job applicants, and they measured introversion/extroversion through the polarity of the words that a candidate uses in their personal blog and then rank them according to the needs of the recruiter. The extroversion score was calculated using the LIWC model developed by Tausczik & Pennebaker (2010). On the other hand, Tandera et al. (2017) developed a personality classifier for Facebook users using deep learning techniques and the Big Five personality model, achieving 74.14% of accuracy. A study by Ortigosa et al. (2014) shows their success finding patterns in Facebook interactions of people with similar personalities based on the Big Five Alternative model. The before mentioned research studies personality prediction using social media and a personality model. This collection of previous work suggests that a machine learning classifier is a viable tool to find a user personality.

Complete Chapter List

Search this Book: