中央研究院 資訊科學研究所

Recently the widespread utilization of social influences (such as viral marketing and product promotions) has received significant attentions. In this talk, we are going to introduce two of our achievements on this field. The first work incorporates the monetary concept into influence diffusion to address the revenue maximization problem. Note that one’s adoption on purchase will depends on the pricing of a product and one’s promotion is convincing when he/she really adopts the product. Based on the principles, we then propose a strategic algorithm, referred to as Algorithm PRUB, to derive the best pricing and the seeded customers for revenue maximization.  

In the second work, we address the existence of interplay between social influence strength and personal interest intensity, and propose new influence diffusion model for personalized promotion through social influence. For this problem, a novel Algorithm ISES is introduced to identify seeded customers in the most cost-effective order.

For both works, experimental results on real social networks demonstrated the effectiveness of the proposed models and approaches, showing great potential of intelligent marketing over social networks.