摘要
研究和应用加权随机森林算法可以有效地解决频分双工—长期演进(FDD-LTE)网络中上行干扰数据存在不平衡的问题,是提高上行干扰分类准确率的有效方法。文章针对测量报告(MR)数据中的上行平均干扰电平,建立了一种基于加权随机森林的上行干扰分类模型,并设置了类权重参数,对比分析了决策树、随机森林和加权随机森林等3种算法的分类效果。结果表明:加权随机森林能够提升不平衡数据中数量较少类的分类正确率,其互调干扰和阻塞干扰的分类正确率分别达到73.91%和96.67%;在不平衡的FDD-LTE上行干扰分类中,加权随机森林能够取得优于决策树和传统随机森林的结果,其分类正确率达到96.22%,而运行时间仅有0.98 s。
The research and application of the weighted random forest algorithm can effectively solve the problem of unbalanced uplink interference data in the Frequency Division Duplexing-Long Term Evolution(FDD-LTE)network.It is an effective method to improve the accuracy of uplink interference classification.Aiming at the uplink average interference level in the Measurement Report(MR)data,a classification model of uplink interference based on weighted random forest is established,and the class weight parameters are set.Then the classification effects of decision tree,random forest and weighted random forest are compared.The results show that the weighted random forest can improve the problem of the classification accuracy of less classes in unbalanced data.The classification accuracy rate of intermodulation interference and blocking interference in fewer classes reaches to 73.91%and 96.67%,respectively.In the unbalanced FDD-LTE uplink interference classification,the weighted random forest can achieve better results than decision trees and traditional random forests.The classification accuracy rate is as high as 96.22%,and the running time is only 0.98 s.
作者
许鸿奎
李鑫
邵星
姜彤彤
宫淑兰
XU Hongkui;LI Xin;SHAO Xing;JIANG Tongtong;GONG Shulan(School of Information and Electrical Engineering,Shandong Jianzhu University,Jinan 250101,China;Shandong Key Laboratory of Intelligent Buildings Technology,Jinan 250101,China)
出处
《山东建筑大学学报》
2020年第6期23-29,共7页
Journal of Shandong Jianzhu University
基金
山东省重大科技创新工程项目(2019JZZY010120)。
