基于加权深度森林的电力调度数据动态挖掘  

Dynamic mining of power dispatching data based on weighted deep forest

作　　者：陈凯阳 骆国铭 CHEN Kaiyang;LUO Guoming(Foshan Power Supply Bureau of Guangdong Power Grid Co.,Ltd.,Foshan 528000,China)

机构地区：[1]广东电网有限责任公司佛山供电局,广东佛山528000

出　　处：《电子设计工程》2025年第8期116-119,124,共5页Electronic Design Engineering

基　　金：广东电网有限责任公司科技项目(202000540023)。

摘　　要：为了解决无法准确挖掘出电力数据中存在的联系,确定隐藏信息的问题,基于加权深度森林提出了电力调度数据动态挖掘方法。聚合和标准化处理信息,并对其进行调试,通过确立采集点的映射关系配置端口参数,建立采集通道,利用I/O点特性确立采集数据点,完成电力调度数据采集。计算林尼指数,并将采集结果输入到4个深度森林中进行训练。通过计算交叉验证因子得到森林模型预测概率,检测森林权重因子,确定数据阈值函数,实现动态挖掘。实验结果表明,基于加权深度森林的电力调度数据动态挖掘方法可以实现平稳训练,最大收敛轮次为30轮,能够快速精准地发现电力调度数据之间的内在联系,确定隐藏信息。In order to solve the problem of accurately mining the connections and identifying hidden information in power data,a dynamic mining method for power dispatch data is proposed based on weighted deep forest.Aggregating and standardizing information processing,debugging it,configuring port parameters by establishing the mapping relationship of collection points,establishing collection channels,and utilizing the characteristics of I/O points to establish collection data points to complete power dispatch data collection.Calculate the Linny index and input the collected results into four deep forests for training.By calculating the cross validation factor,the prediction probability of the forest model is obtained,the forest weight factor is detected,the data threshold function is determined,and dynamic mining is achieved.The experimental results show that the dynamic mining method for power dispatch data based on weighted deep forest can achieve stable training with a maximum convergence round of 30 rounds,and it can quickly and accurately discover the internal connections between power dispatch data and determine hidden information.

关 键 词：加权深度森林 电力调度 调度数据 动态挖掘 林尼指数 

分 类 号：TN02[电子电信—物理电子学]