武汉理工大学学报(信息与管理工程版)
武漢理工大學學報(信息與管理工程版)
무한리공대학학보(신식여관리공정판)
Journal of Wuhan University of Technology (Information & Management Engineering)
2015年
5期
602-606
,共5页
符号网络%共同邻居%边值预测
符號網絡%共同鄰居%邊值預測
부호망락%공동린거%변치예측
signed network%common neighbor%link prediction
针对社会网络中存在的正负二元边值关系,基于共同邻居指标法在识别社会网络符号边值问题中的优势,提出了一种符号网络下的边值预测方法(ICN-Predict)。该符号网络边值预测方法有效结合了节点符号密度属性和网络拓扑结构特征,避免了共同邻居法预测选值敏感性问题。通过实验仿真发现,ICN-Predict预测方法扩大了符号网络边值预测的适用面,提高了边值预测精度,同时表明进一步提高预测精度的关键在于提高负值边的预测准确率。
針對社會網絡中存在的正負二元邊值關繫,基于共同鄰居指標法在識彆社會網絡符號邊值問題中的優勢,提齣瞭一種符號網絡下的邊值預測方法(ICN-Predict)。該符號網絡邊值預測方法有效結閤瞭節點符號密度屬性和網絡拓撲結構特徵,避免瞭共同鄰居法預測選值敏感性問題。通過實驗倣真髮現,ICN-Predict預測方法擴大瞭符號網絡邊值預測的適用麵,提高瞭邊值預測精度,同時錶明進一步提高預測精度的關鍵在于提高負值邊的預測準確率。
침대사회망락중존재적정부이원변치관계,기우공동린거지표법재식별사회망락부호변치문제중적우세,제출료일충부호망락하적변치예측방법(ICN-Predict)。해부호망락변치예측방법유효결합료절점부호밀도속성화망락탁복결구특정,피면료공동린거법예측선치민감성문제。통과실험방진발현,ICN-Predict예측방법확대료부호망락변치예측적괄용면,제고료변치예측정도,동시표명진일보제고예측정도적관건재우제고부치변적예측준학솔。
The social network contains positive and negative edge relations.Based on the common neighbor index method in the recognition of social network signed edge, a kind of signed network link prediction method ( ICN-Predict) was proposed. This method combines the signed density property of the node and the network topology characteristic effectively in order to avoid the problem of selecting sensitivity values in common neighbor method.From the experimental results, ICN-Predict method ex-pands the application of the signed network prediction and improves the prediction accuracy.Meanwhile, it shows that the key point of high accuracy is the predictive ability of negative edge sign.