Science and Engineering Labs (SEL), Room 4211
University of Illinois at Chicago.
Lab phone: (312)-413-5035
E-mail: hhabib3 "AT" uic "DOT" edu
My Ph.D. advisor is Dr. Tanya Y. Berger-Wolf .
Here is my detailed CV.
Algorithmic modelling and analysis of networks and graphs, Machine Learning, Mathematical Modelling of Epidemic and Diffusion processes, Graph mining, and Graph theory.
Dynamic networks generative model. SIAM workshop on network science
Affect of network structure on influence maximization in dynamic networks SIAM workshop on network science
Working for influence: network density and influential individuals. DaMNet 2011.
A Social Networks Approach to Sheep Movement and Leadership. ASNA 2010.
Finding Spread Blockers in Dynamic Networks. LNCS 5498, Start No. 55.
Finding Spread Blockers in Dynamic Networks. SNA KDD 2008.
Graph Theoretic Measures for Identifying Effective Blockers of Spreading Processes in Dynamic Networks. MLG ICML 2008.
The Impact of Structural Changes on Predictions of Diffusion in Networks. ADN ICDM 2008.
Maximizing the Extent of Spread in a Dynamic Network. DIMACS technical report 2007.
Dynamic Betweenness Centrality.
Working for influence: network density and influential individuals.
The problem of finding most influential individuals, or the largest spreaders, in networks has been occupying social scientists, physicists, biologists, epidemiologists, and, lately, computer scientists. Under most even simple models of a spreading process, the question of finding a set of top spreaders turns out to be NP-complete but approximable by a greedy algorithm. Yet, even that greedy algorithm relies on stochastic simulations that can be quite time consuming. In this work we investigate the connection between the network density and the ease of finding the set of top spreaders. We show empirically on both synthetic and real data that for sparse networks simple heuristics like choosing high degree nodes work well, while for really dense networks any randomly chosen set of individuals will perform comparably. There exists, however, an intermediate density region were more sophisticated computational work is needed to find the set of top spreaders. Surprisingly, however, even there the heuristic approaches perform better than the greedy algorithm.