Deep Learning Based Fusion Strategies for Personality Prediction

Automated personality trait detection from text data has emerged and gained a great deal of attention in the subject area of affective computing and sentiment analysis. Most previous work has focused on features engineering such as linguistic styles and psycholinguistic databases which have correlations with personality. Recently, natural language processing has been affected significantly with transfer learning based on feature extraction and fine-tuning pre-trained language models. We propose a new deep learning-based model for personality prediction and classification using both data and classifier level fusion. The model gets benefit from, transfer learning in natural language processing through leading pre-trained language models namely Elmo, ULMFiT, and BERT. The proposed model demonstrates the powerfulness of the introduced method to be a promising personality prediction model. When evaluating the proposed method, results show a competitive and significant accuracy enhancement of about 1.25% and 3.12% in comparison to the most recent results for the two gold standard Essays and myPersonality datasets for personality detection. (c) 2022 THE AUTHORS. Published by Elsevier B.V. on behalf of Faculty of Computers and Information, Cairo University. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).