The project is trying to find the most influential group of nodes in a huge social network.（SUSTech AI lab3） （1）In the ISE problem, I use both LT and IC models. The IC compares the randomly generated values of the nodes with the values of the edges, and is activated if it is less than the edge. The LT model adds the adjacent edges of the node's activated parent nodes, and if the weight value is greater than the valve randomly generated by the node, the modified node is activated. Both algorithms use breadth-first search
（2）In the IMP question, I will calculate the greedy algorithm based on the marginal benefit, and then compare it with the greedy algorithm using the CELF++ optimized algorithm