Target influence maximization problem under avoiding unexpected users interruption in online social network

Abstract

The effectiveness of viral marketing lies in the fact that it can activate a “chain-reaction” driven by word-of-mouth. This paper defines unexpected users’ interruption as the unintentional activation of impedimental users who disseminate negative evaluations. Such interruptions may arise during the process of selecting seed nodes intended to maximize influence spread among target users and foster brand awareness. Furthermore, it is typical to tailor promotions for specific user groups. For instance, consider an event organizer who shares an event advertisement on a social platform, aiming to capture the interest of a large local audience. To solve this problem, we devise positive and target influence maximization under the impedimental influence problem (PTIM), aiming to initiate from a set of influential seed users such that eventually, they will trigger the largest positive influence spread among target users. We prove that the objective function is neither submodular nor supermodular. Then, we decompose it into the difference between two submodular functions (DS decomposition). To solve the maximization problem of DS decomposition under cardinality constraint, we provide an explicit DS greedy algorithm with approximation guarantee. Finally, the experimental simulations on three network data sets demonstrate the effectiveness of our method.