Abstract:With the deepening of population aging, the number of elderly people living alone is increasing rapidly. How to timely detect the abnormal situation of the elderly has become an urgent demand of community management. On the other hand, the development of smart electricity meter technology makes the collection of electricity consumption data more accurate and timelier, which makes it possible to detect the abnormal electricity consumption of the elderly in real time. But how to analyze the massive power consumption data has become a difficulty. Therefore, this paper proposes an early warning model using electricity consumption data for the elderly living alone based on sequential pattern mining, which mines the generality and characteristics from electricity consumption data of the elderly, and comprehensively considers external factors such as time and weather to detect anomalies through deep learning. Compared with traditional detection methods, the model proposed in this paper can detect abnormalities more quickly and accurately, provide security for the life of the elderly living alone, and significantly reduce the pressure of community managers. Experimental results on real data sets verify the validity of the proposed model.