【摘要】 影响力最大化问题是信息网络挖掘中的热门研究问题之一，大多数信息网络包含了多种不同类型的节点和连接边，其本质属于异质信息网络，然而以前关于影响力最大化问题的研究大多停留在同质信息网络，它们考虑的节点和连接边类型单一，这与现实的信息网络有所差别。异质信息网络的影响力最大化问题其关键在于如何识别异质信息网络中最有影响力的节点。为了能融合网络中的异质信息并衡量节点影响力，提出了一种基于加权PageRank的异质信息网络影响力最大化算法。该算法保留了网络中所有类型节点和连接边的信息，通过考虑异质信息网络中不同类型节点之间的影响关系来得到节点的最终影响力，从而实现异质信息网络的影响力最大化。该算法能更好地描述节点和连接边的异质性，并在两个真实的数据集上验证了算法的有效性。更多还原

【Abstract】 The influence maximization problem is one of the hot research problems in information network mining. Most information networks contain a variety of different types of nodes and connection edges, which belong to heterogeneous information networks in nature. However, previous studies on influence maximization mostly stay in homogeneous information networks, which consider a single type of nodes and connection edges and is different from the reality of information networks. The key to influence maximization of heterogeneous information network lies in how to identify the most influential nodes in heterogeneous information network. In order to fuse heterogeneous information in the network and measure the influence of nodes, we propose a weighted PageRank based influence maximization algorithm for heterogeneous information networks, which calculates the influence of nodes through the influence relationship between different types of nodes. The proposed algorithm can better describe the heterogeneity of nodes and edges, of which the validity is verified on two real data sets.更多还原