Limited_IM

The is our implementation of our work entitled "Neighborhood Matters: Influence Maximization in Social Networks with Limited Access".
We proposed both non-adaptive and adaptive algorithms to maximize the influence spread. Specifically, in the non-adaptive case, a two-stage coordinate descent framework is designed to decide the discount allocation. In the adaptive case, a series of algorithms are devised with performance guarantee. Notably, our best algorithm achieves greater approximation ratio than previous work, even under our harsher settings.