Multi-topical authority sensitive influence maximization with authority based graph pruning and three-stage heuristic optimization

Influence maximization is closely related to the interesting topic of users. However, it is often ignored by most previous researches. Even though topics are considered by some previous researches, they neglect users' different authority on different topics. Only one work researches topic authority sensitive influence maximization for a given topic, but its time efficiency is low. What's more, messages usually include more than one topic. In order to solve these problems, we propose a new Multi-Topical Authority sensitive Independent Cascade model (MTAIC), namely the Multi-Topical Authority sensitive Greedy algorithm (MTAG) optimized by the Authority Based Graph Pruning (AGP) and Three-stage Heuristic Optimization Strategy (THOS). A new metric, Influence Spread of seed set on Multi-Topics (ISMT), is put forward to measure the influence spread of seed set considering multi-topics information propagation simultaneously. We do extensive experiments to compare our algorithm with other baseline algorithms on two real-world datasets. Experimental results show that ISMT is an effective measure of influence maximization considering multi-topic authority. The experimental results also demonstrate the effectiveness of MTAIC model and MTAG-AGP, THOS-MTAIC algorithms in terms of ISMT and running time.

Automated summary

In this study, a new Multi-Topical Authority sensitive Independent Cascade model is proposed and compared with other algorithms on real-world datasets, demonstrating its effectiveness.