Automated generation method of recommendation for effective energy utilization as a HEMS service

Smart Grids and Home Energy Management System (HEMS) have been propagated by energy liberalization, and there is a demand for services, which are based on analysis of energy consumption data. For instance, a recommendation on effective utilization of home appliances in order to reduce power consumption. However, it is computationally expensive to analyze data in order to provide an energy-saving handbook, which recommends low-carbon life and is written in a natural language. This kind of service is called recommendation service. Existing automated recommendation services are constrained by the range of data usage, especially when using local information, such as status of surroundings, weather, residents' behavior, etc. The proposed method of automated generation of recommendation considers this background knowledge and information by using clustering methods. The result of a questionnaire which compared a handmade recommendation with the proposed fully-automated recommendation showed that 80% of the residents selected the automated recommendation because of its appropriateness.