基于组合优化LOWESS的电能量数据.doc

基于组合优化LOWESS的电能量数据 缺失处理方法 摘要：针对实际电能量数据的统计分布特性，考虑到均值替代等通常方法对电能量数据缺失的处理效果欠佳，LOWESS模型的估计偏差受限于其给定的窗宽和拟合阶数，提出一种基于预测误差最小化的参数组合优化LOWESS回归模型的电能量数据缺失自动处理方法，通过对比固定窗口和阶数在非平稳的电能量数据上的预测效果，研究参数组合优化LOWESS模型的准确性、适应性以及相对优势性。通过实际数据验证，模型能适应电能量数据不同数据分布、不同缺失比例等情况，预测准确率高，具有一定的实用参考价值。 关键词：缺失数据处理电力计量LOWESS回归组合优化 中图分类号：TM933 文献标识码： 文章编号：1001–1390（201）0–0000–00 Missing electricity data processing method based on combined optimization LOWESS Chen Jun, Long Dong,Yang Zhou,Wei Xingqiu (Electric Power Research Institute of Guangxi Power Grid Company, Nanning 530023, China) Abstract: According to the statistical distribution characteristics of the actual electric energy data, and considering the treatment effect of mean substitution method on energy loss data is usually not satisfactory, the estimation error of LOWESS model is limited by its given window width and fitting order, a LOWESS regression model of the electric energy data deletion optimization based on prediction error minimization parameters automatic processing method is proposed in this paper. By comparing the fixed window and order number in the non-stationarity of the electric energy data on the prediction effect and study the accuracy, adaptability and comparative advantage of parameters optimization LOWESS model. Through the actual data validation, the model can adapt to different data distribution of electric energy data, different loss ratio and so on, and the prediction accuracy is high, which has certain practical reference value. Keywords: missing electricity data, electric energy metering, LOWESS regression, combinatorial optimization 0　引 言 电能计量的智能化、系统化、模块化是电力系统计量技术发展的趋势[1]。由于电能计量设备本身的故障异常 [2]、传输信号不稳定等原因，造成计量自动化终端采集回来的电量计量数据缺失等问题，影响后续电能量数据在计量实际业务的分析、应用。文献[3-6]分别介绍了均值替代、空间自回归、多项式插补、线性插值等通常的数据缺失值处理方法，但当电能量数据不平稳或者缺失偏态情况严重时，通常处理方法的预测误差变得不可控，缺值处理效果欠佳。黄喻[7]、董芳英[8]、苏连成[9]等在地理、基因、材料等领域扩展应用了局部加权回归散点平滑法（Locally Weighted Scatter Plot Smoothing，LOWESS）进行数据缺失处理，并且得到了较好的结果，但现有的LOWESS回归模型其准确性受限于给定的窗口和多项式的阶数。Jae[