摘要
Abstract
Affected by the diversity of group electricity behaviors,the differentiation between different attribute behaviors is se-vere,and it is difficult to classify the diversity characteristics of group electricity consumption behaviors,resulting in a large prediction error.Based on the above reasons,this paper plans to construct a bivariate multi-layer Bayesian prediction model for group electricity consumption behaviors.By introducing the concept of bivariate,analyzing the diversity of group electricity consumption behaviors and dividing the characteristics of group electricity consumption behaviors,this paper constructs a multi-layer Bayesian prediction model,uses bivariate multi-layer Bayesian model to calculate the prior distribution of electricity con-sumption behaviors characteristics,and derives the distribution parameters for output modeled predictions.The experimental results show that the recognition rate can be maintained at more than 87％on the four difficulty indicators,the trend of the pre-diction response time change curve is stable,the prediction accuracy,response speed,and recognition ability meet the predic-tion requirements of group electricity consumption behaviors,which has high application and promotion value.关键词
群体用电行为/双变量模型/多层贝叶斯预测模型/分布参量Key words
group electricity consumption behavior/bivariate model/multi-layer Bayesian prediction model/distribution param-eter分类
信息技术与安全科学
