Predicting interval time for reciprocal link creation using survival analysis
Predicting interval time for reciprocal link creation using survival analysis
复制标题
DOI：
10.1007/s13278-018-0494-1
复制
发表时间：
2018-03
影响因子：
2.8
通讯作者：
Vachik S. Dave;M. Hasan;Baichuan Zhang;Chandan K. Reddy
中科院分区：
文献类型：
--
作者：
Vachik S. Dave;M. Hasan;Baichuan Zhang;Chandan K. Reddy
大多数有向社交网络，如Twitter、Flickr和Google+，都存在互惠利他主义，这是一种社会心理学现象，它驱使一个顶点与另一个顶点建立互惠链接，而另一个顶点已经建立了指向前者的有向链接。在现有的工作中，科学家们已经预测了建立相互联系的可能性-一项被称为“相互联系预测”的任务。然而，一个同样重要的问题是确定第一个链接（也称为准社会链接）的创建与其相应的相互链接之间的间隔时间。现有的研究工作还没有考虑解决这一问题，这也是本文的重点。预测互逆链接间隔时间是一个具有挑战性的问题，原因有两个：第一，缺乏有效的特征，因为众所周知的链接预测特征是为无向网络和二进制分类任务设计的;因此，它们不能很好地用于间隔时间预测;第二，存在过度等待链接（即，在观察期内没有形成相互联系的准社会联系）使得传统的监督回归方法不适合于这种数据。在本文中，我们提出了一个解决方案的倒数linkinterval时间预测任务。我们将这个问题映射到生存分析任务，并通过对真实世界数据集的大量实验表明，生存分析方法在解决倒数间隔时间预测方面优于传统回归，基于神经网络的模型和支持向量回归。
The majority of directed social networks, such as Twitter, Flickr and Google+, exhibitreciprocal altruism, a social psychology phenomenon, which drives a vertex to create a reciprocal link with another vertex which has created a directed link toward the former. In existing works, scientists have already predicted the possibility of the creation of reciprocal link—a task known as “reciprocal link prediction”. However, an equally important problem is determining theinterval timebetween the creation of the first link (also called parasocial link) and its corresponding reciprocal link. No existing works have considered solving this problem, which is the focus of this paper. Predicting the reciprocal linkinterval timeis a challenging problem for two reasons: First, there is a lack of effective features, since well-known link prediction features are designed for undirected networks and for the binary classification task; hence, they do not work well for theinterval timeprediction; Second, the presence ofever-waitinglinks (i.e., parasocial links for which a reciprocal link is not formed within the observation period) makes the traditional supervised regression methods unsuitable for such data. In this paper, we propose a solution for the reciprocal linkinterval timeprediction task. We map this problem to a survival analysis task and show through extensive experiments on real-world datasets that survival analysis methods perform better than traditional regression, neural network-based models and support vector regression for solving reciprocalinterval timeprediction.