Contemporary e-Government designs incorporate machine learning enabling machine-to-machine communication. In such a scenario, machines are able to make decisions by analyzing huge sets of data and policies to make coherent decisions. Traditional machine learning (ML) models have been incorporated into Large Language Models (LLMs) to mimic decision-making o individuals in public service delivery value chains. This paper explores a non-traditional orientation to machine learning using a highly statistical approach8. The paper provides baseline statistical models which could be incorporated into LLMs in e- Government platform and solution designs. It is hoped that statistical approaches to ML using LLMs can open future research and practical design of e-Government solutions. Given that contemporary e-Government systems focuses on data-centric designs, it is important to explore data-centric designs in creating public service ecosystems based on data.
Bwalya Kelvin Joseph (Thu,) studied this question.