Machine Learning for User Learning

Abstract

To guide users through their business application’s functionality requires an intelligent digital assistance system to adapt to the user’s stage of expertise. Drawing on event segmentation theory and knowledge space theory, we propose to model the users’ domain specific knowledge and their learning process dynamically in the interaction between system and user. In the support process, the system retrieves the support content that matches the user’s knowledge state from a hierarchically organized case base. Using case-based reasoning as a psychologically inspired machine learning method facilitates incorporating the user’s feedback in the interaction: the system continuously updates its user model to learn how to support the user most efficiently and effectively.