融合节点重要性的通信系统无特征网络链路预测方法

Characterisless Network Link Prediction Method of Communication System Incorporating Node Importance

通信系统的无特征网络链路中,由于忽略了节点的特征属性,导致预测评估结果的曲线下面积(Area Under Curve,AUC)值较低。针对上述现象,提出融合节点重要性的通信系统链路预测方法。提取无特征网络中节点的局部特征,构建节点的时间序列数据。计算每个节点的重要性,利用节点的重要性和时间序列数据,通过特定的算法,获取每个节点的预测值,从而实现无特征网络链路的预测。实验结果表明,该方法预测评估结果的AUC值较高,能够更准确地预测网络中的链路连接。

In featureless network links of communication systems,the neglect of node feature attributes leads to lower Area Under Curve(AUC)values in prediction and evaluation results.Propose a communication system link prediction method that integrates node importance in response to the above phenomenon.Extract local features of nodes in featureless networks and construct time series data of nodes.Calculate the node importance of each node,utilize the node importance and time series data,and use specific algorithms to obtain the predicted values of each node,thereby achieving the prediction of featureless network links.The experimental results show that the method has a higher AUC value for predicting evaluation results and can more accurately predict link connections in the network.