Predicting interaction design patterns for designing explicit interactions in ambient intelligence systems: a case study.

Ambient intelligence (AmI) focuses on supporting people by designing sensitive and responsive environments to context through implicit and explicit interactions. Explicit interactions in AmI systems have requirements specific to making interactions robust, smooth, intuitive, and reliable. Based on requirements, the designers can detect and eliminate faults from the beginning of the design process and understand the users' needs and demands. This work presents a UIPatternM model for predicting interaction design patterns from processing text-based requirements through machine learning algorithms. We evaluate the predictions of our proposal. We also present a case study with professional designers who evaluated the UIPatternM recommender predictions according to a set of design-level requirements that emulate everyday needs. Our participants performed a set of tasks based on scenarios, and we evaluated the participants' using effectiveness, efficiency, and satisfaction as performance metrics. Applying the UIPatternM model helped to endorse the conception and refinement of user interface design for explicit interaction in AmI systems. [ABSTRACT FROM AUTHOR]