Leveraging Machine Learning Algorithm for Predicting Personality Traits on Twitter

Social networks' rapid rise has brought with it many ways of understanding individual interactions within these digital ecosystems and the language employed within. These methodologies include machine learning algorithms, information sources and feature-based approaches - offering a broad array of techniques currently available. This research endeavor's primary objective is to explore the predictability of personal characteristics among Facebook users based on components and metrics drawn from Big 5 personality model. Researchers utilized the Personality Research dataset to investigate social media networks and language traits relevant to interpersonal personality interactions. Four machine learning algorithms were extensively evaluated before performing correlation analyses between each set feature and user attributes. Even with similar datasets tested, XGBoost-classifiers consistently outshone baseline methods in terms of prediction accuracy. Researchers achieved an 85.3 percent accuracy rate. Social network analysis features sets proved instrumental in increasing accuracy by 89.5 percent for personality attribute predictions; this demonstrated their strong predictive capabilities as the central feature in this research to enhance personality prediction.