Hybrid Learning Approach Toward Situation Recognition and Handling

Abstract We propose a novel hybrid learning approach to gain situation awareness in smart environments by introducing new identifier that combines an expert system and machine approach. Traditionally, systems approaches have been widely used independently detect ongoing situations as the main functionality various domains. Expert lack adapt each user are expensive design based on setting. On other hand, fail challenge of cold start making explainable decisions. Using both these enables use user’s feedback capture environmental changes while exploiting initial knowledge solve mentioned challenges. decision trees templates core structure interpret sensor data. To evaluate proposed method, we generate human-annotated dataset simulating environment. Our experiments show superior results compared with preserving system’s interpretability.