| 黄寿孟1，2.一种基于节点路径信息相似性的预测方法[J].海南师范大学学报自科版,2022,35(1):25-30 |
| 一种基于节点路径信息相似性的预测方法 |
| A Link Prediction Method Based on Similarityof Node Path Information |
| |
| DOI：10.12051/j.issn.1674-4942.2022.01.004 |
| 中文关键词: 复杂网络 链路预测 节点信息 |
| 英文关键词: complex networks link prediction node information |
| 基金项目:海南省高等学校科学研究一般项目（Hnky2021-51） |
|
| 摘要点击次数: 645 |
| 全文下载次数: 833 |
| 中文摘要: |
| 链路预测计算是在复杂网络分析任务中最重要和最具挑战性的任务之一，能根据网络
中现有的链接预测缺失的链接并广泛应用于多种学科领域，包括社会网络分析、推荐系统和生物
网络等。文中提出一种基于路径节点信息相似性的预测方法，该预测方法是利用节点共有的特征
信息来推测下一个相关的路径节点信息，从而优化现有的基于路径预测方法。首先，由于网络中
一对可达节点之间的最短距离始终是唯一的，所以通过合并本地路径节点信息计算出网络中所有
节点对之间最短路径中的路径最大长度，从而提高节点路径相似性的准确性；接着，以类似的方式
计算遍历所有未连接的节点对，计算它们预测分配的权重得分，得到将来可能会链接起来的节点
相似性得分，即是这些节点中某两个节点的最短路径；最后，采用来自不同领域的真实世界数据集
进行实验实证。结果表明，与现有的最先进的预测方法相比，该方法具有较高的预测精度。 |
| 英文摘要: |
| Link prediction computing is one of the most important and challenging tasks in complex network analysis. It
predicts missing links based on existing links in networks, and is widely used in various scientific fields, including social
network analysis, recommendation systems and biological networks. In this paper, a prediction method based on the similar⁃
ity of path node information was proposed. The prediction method used the common feature information of nodes to predict
the next related path node information, so as to optimize the existing path prediction method. As the shortest distance be⁃
tween a pair of nodes in network is always unique, by combining the local path node information, calculating the shortest
path between all nodes in network to the maximum length of path and to improve the accuracy of node path similarity, and
then calculating traverse all the unconnected nodes in a similar way and weighted score of distribution, the similarity score
of nodes that may be linked in the future was obtained, that is, the shortest path of two nodes in these nodes. Finally, the re⁃
al world data sets from different fields were used for experimental verification. The results show that the method has higher
prediction accuracy, compared with the most advanced prediction methods. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
