无线协作中继网络多层不良数据辨识方法

Multi-Layer Bad Data Identification Method for Wireless Cooperative Relay Network

针对传统方法没有对不良数据进行特征点提取,导致总体准确率低、召回率低和F1比值低的问题,提出无线协作中继网络多层不良数据辨识方法。利用决策树对传统数据传输通道进行优化,将不良数据平均分散到各个通道,在信号模型的基础上构建网络多层不良数据特征点提取模型,提取网络多层不良数据特征点。利用COPS算法对不良数据特征向量进行聚类,将其归一化后增量,获取聚类划分结果,同时运算出对应的聚类指标,聚类结果中平衡类内紧凑和类间分离的点就是最优解,即指标中的最小值就是最优解,实现网络多层不良数据辨识。实验结果表明,所提方法的总体准确率高、召回率高和F1比值高,说明该方法具有应用价值。

Traditional methods ignore the feature points extraction of bad data, resulting in low accuracy, recall and F1 ratio. Therefore, this paper proposes a multi-layer bad data identification method for wireless cooperative relay networks. Based on the decision tree, the traditional data transmission channel was optimized. Bad data were evenly distributed to each channel. According to the signal model, the feature points extraction model of network multi-layer bad data was established to extract the feature points of network multi-layer bad data. Through COPS algorithm, the feature vectors of bad data were clustered, normalized and incrementally processed for obtaining the clustering results. Meanwhile, the corresponding clustering indexes were calculated to obtain the optimal solution, thus achieving the identification of multi-layer bad data in the network. The results show that the method has high overall accuracy, recall, F1 ratio and outstanding application value.