基于协同过滤推荐的电价计划智能推荐算法  

Intelligent recommendation algorithm for electricity price plan based on collaborative filtering recommendation

作　　者：钟华赞 包新晔[1] 杜杰 ZHONG Hua-zan;BAO Xin-ye;DU Jie(Production Technology Department,China Southern Power Grid Limited Company,Guangzhou 510623,China;College of Electrical Engineering,Tianjin University,Tianjin 300011,China)

机构地区：[1]中国南方电网有限责任公司生产技术部,广东广州510623 [2]天津大学电气工程学院,天津300011

出　　处：《计算机工程与设计》2021年第12期3525-3533,共9页Computer Engineering and Design

基　　金：国家自然科学基金项目(61872265、61402325);南方电网公司科技基金项目(ZBKJXM20170230)。

摘　　要：在智慧电网中,电力公司可以主动推荐定制的售电方案给潜在用户,但现有的推荐算法存在着精确度不高、方案不合理等缺点。为解决以上问题,基于协同过滤策略,开发一种电力计划推荐方案。通过提供一些容易获得的家电产品数据,对居民用户进行不同方案的预测评级,为用户选择合适的方案和合理的电价。在实验阶段,通过不同的数值实验评价该方法的性能,实验结果表明,EPR算法在推荐结果的准确性上优于其它策略。In the smart grid,power companies can actively recommend customized power sales solutions to potential users,but the existing recommendation algorithms have disadvantages such as low accuracy and unreasonable solutions.To solve the above problems,a power plan recommendation scheme was developed based on the collaborative filtering strategy.By providing some easy-to-obtain household electrical appliance product data,forecast ratings on different schemes were carried out for residential users,and suitable schemes and reasonable electricity prices were selected for users.In the experimental stage,the performance of the method was evaluated through different numerical experiments.The results show that the EPR algorithm is better than other strategies in the accuracy of the recommendation results.

关 键 词：机器学习 协同过滤 智慧电网 推荐系统 电费开销 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]