Abstract:
The traditional influence maximization algorithm ignores the commercial income problem in the process of viral marketing. In real marketing, businesses pay more attention to how to use a fixed budget to select seed sets in a reasonable time to maximize marketing benefits. To solve this problem, an efficient heuristic algorithm is proposed. Firstly, define marginal cost performance to measure the importance of user nodes. Secondly, the node selection result of greedy algorithm is analyzed as a self-consistent sequence, and MCPR (Marginal Cost Performance Ranking) algorithm is proposed, which iteratively approximates an approximate self-consistent sort to pursue the effect of greedy algorithm. Finally, the cost-effective forward allocation strategy is used to estimate the marginal cost-effective of nodes and accelerate the algorithm iteration. A large number of experiments were carried out on three real social networks, and the results show that MCPR can achieve similar results with the greedy algorithm, but the algorithm efficiency is much higher than that of the greedy algorithm.