In recent years, social media has emerged as a rich data source for diverse research applications, including natural language processing (NLP) and machine learning (ML) projects. Among these, predicting personality traits from social media content has gained significant attention due to its potential to reveal fundamental psychological attributes that influence life outcomes such as academic success, job performance, and social status. Traditional methods for personality assessment, such as self-questionnaires and expert observation, are costly, time-consuming, and prone to errors. This challenge is particularly evident in hiring, where evaluating the personalities of thousands of applicants is crucial to ensure alignment with job requirements. Our study focuses on addressing this challenge by employing a range of ML algorithms and multi-label classification techniques (Classifier Chain, Binary Relevance, and Label Powerset) to extract Big Five personality traits (openness, conscientiousness, extraversion, agreeableness, and neuroticism) from LinkedIn profiles. Results demonstrate the superior performance of ensemble methods, particularly the Classifier Chain technique with Linear SVM, which achieved a Hamming Loss of 0.17 and an F1-Score of 0.80. These findings highlight the potential of integrating social network data and advanced ML techniques to streamline personality assessment, offering valuable insights for optimizing recruitment processes.

Keywords: Machine-Learning (ML), Supervised-Learning (SL), Multi-Label Classification, Big-Five Model, Personality Traits.