Online Influence Maximization under Decreasing Cascade Model

TL;DR

This paper introduces the decreasing cascade (DC) model for online influence maximization, accounting for market saturation effects, and proposes the DC-UCB algorithm with proven regret bounds and strong empirical performance.

Contribution

It generalizes the independent cascade model to include decreasing influence probabilities and develops a new algorithm with theoretical guarantees for this setting.

Findings

DC-UCB achieves low regret bounds similar to IC models

Experimental results confirm the effectiveness of DC-UCB on synthetic and real data

The decreasing cascade model better captures market saturation phenomena

Abstract

We study online influence maximization (OIM) under a new model of decreasing cascade (DC). This model is a generalization of the independent cascade (IC) model by considering the common phenomenon of market saturation. In DC, the chance of an influence attempt being successful reduces with previous failures. The effect is neglected by previous OIM works under IC and linear threshold models. We propose the DC-UCB algorithm to solve this problem, which achieves a regret bound of the same order as the state-of-the-art works on the IC model. Extensive experiments on both synthetic and real datasets show the effectiveness of our algorithm.