Interactive Data Acquisition for CBR System Based Smart Home Assistant

Abstract

This research aims to develop a Case-Based Reasoning (CBR) system that recommends services to users in IoT environment. To develop this system, we establish a framework that designs raw data into analyzable information using Function-Behavior-Structure properties. Also, we develop an interactive flow of data acquisition that builds up cases gradually by gathering data through conversational interactions between the system and its user. This research develop a prototype of this system based on simulated cases. Finally, the prototype of this system was evaluated by experts in the field of system design to verify how the service (solution) recommended by system is similar with them. The results of this evaluation showed an agreement of average 54%, but found that there was a big difference from the experts in the specific context. This result implies that it is necessary to improve the context awareness in the reasoning process of this system.