Predictive Models of Typographic Preference in Digital Media

Introduction: This article explores how typography influences user experience in digital environments, highlighting its evolution from the 11th century to the Internet era. Objective: The aim of this research was to examine the psychological impact of fonts, which evoke emotional responses and affect readability, design and user behavior. Methodology: Predictive models, such as regression, classification and time series, are used to analyze typographic preferences, helping designers to optimize digital interfaces. Results: The study simulated data from 1,000 participants, considering variables such as age, gender, educational level and context of use, revealing a predominant preference for Sans Serif typefaces (63.3%), especially in academic reading. The Logistic Regression and SVM models showed a moderate performance (accuracy of 0.627 and 0.634), with better ability to identify preferences for Sans Serif, although with limitations for the minority class (Serif). Conclusion: It was concluded that psychological, cultural and contextual factors significantly influence preferences, highlighting the need to integrate these variables in future models to improve accuracy and personalization in digital design.