Seeing the Forest Through the Traits: A Four-Dimensional Framework for Modeling and Measuring Latent User Traits

Understanding user traits is essential for developing adaptive and personalized systems. However, current methods often depend on lengthy and mentally taxing assessments or on user information that fails to capture the nuances of personality, cognitive abilities, or perceptual skills. This paper introduces a simplified approach to building a user representation by proposing a four-dimensional framework that captures Perceptual, Relational, Exploratory, and Self-Agency dimensions. Our model was derived from a dataset we collected with 305 participants across 16 latent user traits. These traits were chosen from widely used psychometric tools, including assessments for the Big-Five personality traits, Locus of Control, curiosity, and perception. Leveraging predictive modeling, we identified the most informative features for each dimension, resulting in a concise and reliable questionnaire. The reduced questionnaire balances brevity with accuracy, retaining strong classification performance and internal consistency across dimensions. This work advances user modeling by offering both a novel dimensional structure and a transparent, interpretable process for question reduction, paving the way for scalable, user-friendly, trait-based personalization in adaptive systems.