【摘要】：網(wǎng)絡(luò)逐漸成為了人與人之間的主要社交工具,在網(wǎng)絡(luò)中挖掘最有影響力的用戶成為了非常值得關(guān)注的問題。在傳統(tǒng)影響力最大化算法的基礎(chǔ)上提出了一種面向主題耦合的影響力最大化算法,該算法首先分析網(wǎng)絡(luò)中不同主題之間的耦合相似性,在綜合考慮主題之間耦合相似性與用戶對(duì)不同主題偏好的基礎(chǔ)上擴(kuò)展獨(dú)立級(jí)聯(lián)模型,并使用經(jīng)典的貪心算法挖掘最具有影響力的用戶。與不考慮主題耦合的影響力最大化算法相比,所提算法考慮了傳播主題之間的耦合相似性,并且能夠與用戶偏好進(jìn)行更為有效地結(jié)合。最后,實(shí)驗(yàn)表明,相比于經(jīng)典的影響力最大化算法,該算法能夠更為有效地挖掘在特定主題下最具有影響力的種子節(jié)點(diǎn)。
[Abstract]:The Internet has gradually become the main social tool between people, mining the most influential users in the network has become a very important issue. Based on the traditional influence maximization algorithm, a topic coupling oriented influence maximization algorithm is proposed. Firstly, the coupling similarity between different topics in the network is analyzed. On the basis of considering the coupling similarity between topics and users' preference for different topics, the independent cascade model is extended, and the most influential users are mined by the classical greedy algorithm. Compared with the influence maximization algorithm which does not consider the topic coupling, the proposed algorithm takes into account the coupling similarity between propagating topics and is more effective in combining with user preferences. Finally, experiments show that the algorithm can effectively mine the most influential seed nodes under a particular topic compared with the classical algorithm.
【作者單位】： 云南大學(xué)信息學(xué)院;
【基金】：國(guó)家自然科學(xué)基金項(xiàng)目(61262069,61472346,61762090) 云南省自然科學(xué)基金項(xiàng)目(2015FB114,2016FA026) 云南省創(chuàng)新團(tuán)隊(duì) 云南省高�？萍紕�(chuàng)新團(tuán)隊(duì)(IRTSTYN) 云南大學(xué)創(chuàng)新團(tuán)隊(duì)發(fā)展計(jì)劃(XT412011)資助
【分類號(hào)】：TP301.6
，

本文編號(hào)：2422173