| Abstract |
Influence maximisation, i.e., the study of strategically influencing agents that aim at spreading opinions, behaviours, or rumors on social networks, has been extensively investigated in recent work, mostly via variants of the seminal independent cascade model [1]. However, algorithms that provide optimal and nearly-optimal solutions to the problem of triggering an influence cascade, by selecting the optimal seed node sets, often achieve this only if certain assumptions are met. Among these assumptions, the most stringent and common is the availability of complete information about a social network. Not surprisingly, this is usually not the case, especially when dealing with Big Data and offline social networks. |