Use of Machine Learning to Validate an Intelligent Framework to Support Decision Making in the Public Sector

Abstract

Problem: Currently, many public organizations have adopted applications for process automation to avoid repetitive work and produce more efficient results; however, the development of intelligent mechanisms to support complex decision-making is not often observed. In public services, in particular, difficulties may be related to the abundance of data sources and the number of legal norms to comply with. Objective: A formal specification of a framework for the application and service layer suitable for public services with machine learning to support decisionmaking by technology and business experts. Method: In this study, the Design Science Research Methodology (DSRM) method was used, dividing the work into the following stages: (i) identification of the problem and motivation; (ii) definition of the objectives; (iii) planning, design, and development; (iv) demonstrations of the simulations; (v) verification and validation of the experiments; and (vi) communication of results. Interspersed with Domain Engineering (DE) in three stages: (i) Domain Analysis, (ii) Domain Design, and (iii) Domain Implementation. Results: This research was carried out: (i) elicitation of characteristics for an Intelligent Framework for the Public Sector, (ii) execution of Domain Engineering in Public Sector projects to obtain the characteristics, (iii) construction of an architectural model with machine learning by reinforcement, and (iv) instantiation of the framework for its validation using five experimental cases.