Document Type

Article

Publication Title

Journal of Data Science

Abstract

The traditional approach to biographical profiling, predominantly reliant on limited and fragmented datasets, has frequently resulted in superficial personality understandings. This is largely due to an overemphasis on official records and notable events, neglecting the rich tapestry of everyday experiences and personal interactions that significantly shape personalities. To address this shortcoming, this article introduces a multi-disciplinary methodology, The Holistic Archival Personality Profiling Model (HAPPM), which integrates a diverse array of archival materials, including personal correspondences, social media footprints, and family memorabilia. This approach involves digitizing various data forms, including handwritten documents, into machine-readable text, and then semantically classifying this data with biotags, chronotags, and geotags for organization within specific spatial and temporal contexts. Such comprehensive data aggregation establishes a more accurate "space-time continuum" for individuals, enhancing our understanding of their lives. The innovative aspect of HAPPM is the utilization of large language models to "converse" with the data, facilitating a more holistic representation of personalities. Preliminary results from applying HAPPM have shown its efficacy in uncovering previously unknown aspects of individual lives, offering insights into personal beliefs, daily routines, and social interactions. This has been validated through comparative analysis with existing biographical data, revealing a more complete and nuanced understanding of personalities. Therefore, HAPPM marks a significant advancement in personality profiling, capturing not only the grandiose but also the mundane, and offering a comprehensive tool for researchers and historians to explore the full spectrum of human experience.

Publication Date

12-2023

Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

Download

Share

COinS