A STUDY ON ESTIMATING WAKE TIME USING TIME-SERIES ELECTRICITY CONSUMPTION DATA FOR THE CARE OF SINGLE ELDERLY RESIDENTS IN URBAN DWELLINGS

The purpose of this study was to develop a model for estimating wake time using only time-series electricity consumption data and to verify its accuracy. Presence/absence, sleeping/non-sleeping, air conditioner on/off, and electricity consumption data obtained from a survey of 12 single elderly people over a one-year period were used. The model classifies people into two categories: active and inactive, and estimates their wake time from the results of the two classifications. When the two factors that reduced accuracy were excluded, it was possible to estimate wake times within one hour of estimation error on 278 days of the year.