论文部分内容阅读
微博中的意见领袖不仅在社交网络的信息传播中发挥着举足轻重的作用,而且在网络舆情演化中也表现出显著的意见代表性。针对已有的意见领袖挖掘方法仅从复杂网络或者基本图模型来建模发现意见领袖,忽略了意见领袖在具体的话题演化中的意见代表性的问题,提出了基于话题演化的意见领袖发现的方法。该方法首先根据用户之间的交互构建图模型,然后利用寻找中心节点的图论算法挖掘潜在意见领袖,再利用话题演化模型判断潜在意见领袖的演化中心度,最后发现在整体舆情上的具有意见代表性的真实意见领袖。在新浪微博的话题数据集上的试验结果表明,该算法较仅考虑网络模型的意见领袖发现方法更优。
Opinion leaders in Weibo not only play an important role in the information dissemination of social networks, but also display conspicuous representations of opinions in the evolution of online public opinion. Aiming at the problem that existing opinion leaders ’mining method models the opinion leaders only from the complex network or the basic graph model and ignores the representative opinion of the opinion leaders on the specific topic evolution, this dissertation proposes the opinion leaders’ opinion based on topic evolution method. The method first builds the graph model according to the interaction between users, then digs the potential opinion leader by using the graph theory algorithm of looking for the center node, then uses the topic evolution model to judge the evolutionary center of the potential opinion leader, and finally finds the opinion on the overall public opinion Representative of the real opinion leaders. The experimental results on Sina Weibo’s topic dataset show that the proposed algorithm is superior to the opinion leader discovery method that only considers the network model.