Personality Prediction in Business using Artificial Intelligence and Machine Learning

The main goal of this method is to propose the development of a very precise, robust, and scalable system for predicting a personality using the most advanced techniques of deep learning, with a special focus on BERT-based NLP approaches. Differently from how machine learning is usually performed, in which the process is very reliant on carefully designed language features and shallow text patterns, this system would instead use contextualized word embeddings to automatically detect deep semantic meanings, underlying emotional patterns, as well as implicit behavioural features expressed in texts. The goal of this research is to show how the application of transformer architectures is able to enhance the trustworthiness and accuracy of inferring personality traits, especially in relation to the Big Five dimensions of personality, namely Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism. Through their application on business-related corpuses, as well as on different digital communication texts. The proposed research aims to develop a generalized predicting system, scalable enough to process large amounts of unstructured texts, while still showing high predicting power and interpretability. Finally, this work is intending to propose an advanced contribution to the ever-growing computerized psychology community, in the form of an innovative, fully automatic, and precision-driven system for personality analysis.