Abstract:

In recent years, there has been a significant build-up of published research on deploying various machine learning (ML) methods and algorithms in research supporting decisions in higher education institutions (HEIs) processes. On the contrary, a small number of research show the benefits of the developed methodology to improve process management. At the same time, there is mass evidence that leading IT companies worldwide utilize intelligent solutions to solve everything. In this paper, the authors present the integration of an intelligent agent into the HEI information system that optimizes the benefit of the different ML algorithms' usage in answering standard queries that various HEI services search for in the existing data. Users should not be concerned with the method selection, so the agent has to select the proper method. That selection has to be data-driven, meaning the agent has to run various algorithms with the existing data and select the most proper one. However, users usually have personal preferences, meaning they understand some outputs from some methods and algorithms more easily than others. For this reason, the AI agent should have feedback from the user on how understandable the output of particular methods and algorithms is to them and how useful the information obtained as an output is. The agent's performance is tested on the ML.NET solutions and the data from Distance Learning Web Management System (DLWMS) from the Faculty of Information Technologies (FIT), University Dzemal Bijedic of Mostar, Bosnia and Herzegovina.

Keywords:

LMS, intelligent agent, ML methods.