An Integrated Agent-Based Modelling and Machine Learning Framework for Enhancing the Digital Service Access of Minority Ethnic Communities

This paper introduces an integrated modelling framework combining Agent-Based Modelling (ABM) and Machine Learning (ML) to enhance digital service access for Minority Ethnic (ME) communities, specifically drawing on the Protecting Minority Ethnic Communities Online (PRIME) project. The framework leverages ML models to derive unbiased decision-making rules, influencing ABM simulations that explore online service perceptions based on individual attributes and environmental factors. It facilitates comprehensive analysis of agent interactions, policy impacts, and service provider characteristics, aiming to improve online service experiences across health, energy, and social housing domains. By integrating ML with ABM, the framework addresses key challenges in model design, variable handling, and validation, using empirical data and expert knowledge. The framework’s adaptability supports detailed examination of enablers and constraints affecting minority ethnic communities’ online service usage, promoting equitable policy decisions and service delivery improvements. Future extensions include optimization for equitable outcomes and threat modelling for secure online service access.