Problem Formulation for the Development of an Intelligent System for Predicting Future User Needs Based on a Dynamic Multimodal Digital Persona
Problem Formulation for the Development of an Intelligent System for Predicting Future User Needs Based on a Dynamic Multimodal Digital Persona
Vardanyan Hayarpi, Saakyan Rustam
Summary
Key words: reactive agents, proactive agents, multimodal learning, behavioral modeling, dynamic model, user modeling
This paper presents the problem formulation of an intelligent system for predicting future user needs, based on a dynamic multimodal digital persona. The study of the subject domain made it possible to clearly distinguish the key characteristics of reactive and proactive agents, demonstrating that a deep and multi-layered understanding of the user is a necessary precondition for building a predictive system.
In the course of the research, it was substantiated that existing concepts of user profiles and models, being predominantly static in nature, are insufficient for proactive prediction. In addition to these, it is necessary to apply dynamic approaches to behavioral modeling, encompassing the analysis of micro- and macro-level interactions as well as multi-behavior interactions.
The presented analysis demonstrates that the transition from the reactive “request–response” paradigm toward predictive, proactive interaction is achievable, provided that the system possesses a multidimensional, continuously updated digital persona of the user. Since traditional static models do not adapt to the rapid changes of the digital domain, a transition to multilayered behavioral modeling is proposed in order to enhance the accuracy of predictions. This approach holds applied significance in the fields of education, healthcare, information services, and personalized systems.
Future research will focus on the development of a system based on the proposed model and its applied validation using multi-domain data.