User demand response method and device, terminal and storage medium

The invention discloses a user demand response method and device, a terminal and a storage medium. For the user power utilization parameters, users are classified based on a user classification modelof a pre-training model, and the pre-training model is trained by a nearest neighbor classifier; and the user power consumption mode is adjusted for the user according to the user classification category. The considered parameters comprise daily power consumption, weekly power consumption, monthly power consumption, daily average peak power consumption duration and daily peak power consumption time period, the daily power consumption, weekly power consumption and monthly power consumption are considered, abnormal conditions are effectively eliminated, and the classification accuracy is improved. 本发明公开一种用户需求响应方法、装置、终端及存储介质，获取用户用电参数；对于用户用电参数，基于预先训练模型的用户分类模型对用户进行分类，所述预先训练模型由最近邻分类器进行训练；根据用户分类的类别对用户调整用户用电模式。本发明所考虑参数包括日用电量、周用电量、月用电量、每日平均峰值用电时长、每日峰值用电时段，不仅考虑每日的用电量，还考虑周用电量和月用电量，有效剔除异常情况，提高分类准确性。