基于机器学习的用户窃电行为预测

新型智能电表普及后,为了准确检测出电网中的窃电用户,可以结合机器学习的方法.为此,选择了支持向量机、随机森林和迭代决策树3种机器学习中较常用的大数据算法进行分析,通过不断调整试验数据集的大小,对3种算法的效率和准确率进行测试.对比分析结果发现,随机森林算法运行的时间和数据量的大小基本呈线性关系,效率较高,且准确率稳定在86%以上,表现较好.

Prediction of User Stealing Behavior Based on Machine Learning

Accurate detection of the power grid users can be combined with the machine learning method after the popularity of new smart meters.For this purpose,three kinds of machine learning more commonly used in large data algorithm are chosen for analysis:random forest,support vector machine and gradient boosting decision tree.The efficiency and accuracy of the three algorithms are tested by constantly adjusting the size of the test data set.Analysis of the results shows that the random forest algorithm runs in a linear relationship with the amount of time and the amount of data,while the accuracy rate of stability is higher than 86%,with better performances.