PersonaG: A Quinpartite Graph Convolutional Network for Interpretable Personality Recognition from Text

Abstract

Automatic personality recognition from text has wide-ranging applicationsin social media analysis, targeted marketing, and personalized user experiences. As such, a lot of researchers have focused on personality recognition in the last two decades. However, existing methods often rely on only shallow semantic features or psycholinguistic features to capture the semantic information in textual data. We propose PersonaG, a novel approach that integrates psycholinguistic categories with WordNet semantics to address these limitations and construct quinpartite graph representations. Our approach combines semantic relationships with the psycholinguistic categories. Classification is performed using a Dynamic Deep Graph Convolutional Network. Our results on the benchmark Essays dataset outperform recent methods, achieving state-of-the-art performance and demonstrating the superiority of our approach. To conclude, the quinpartite graph enables PersonaG to understand the latent personality patterns from text, making it a comprehensive and effective solution for personality recognition.

Received: 16 October 2024 | Revised: 27 November 2024 | Accepted: 16 December 2024

Conflicts of Interest

The authors declare that they have no conflicts of interest to this work.

Data Availability Statement

The data that support the findings of this study are openly available in Kaggle at https://psycnet.apa.org/doi/10.1037/0022-3514.77.6.1296, reference number [17].

Author Contribution Statement

Mohmad Azhar Teli: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization, Project administration. Manzoor Ahmad Chachoo: Conceptualization, Resources, Supervision, Project administration.