Positive Influence Dominating Set generation in social networks
read more
Citations
Rapid Influence Maximization on Social Networks: The Positive Influence Dominating Set Problem
Influence Maximization with Latency Requirements on Social Networks
Total positive influence domination on weighted networks
Influence Maximization with Latency Requirements on Social Networks
Total Positive Influence Domination on Weighted Networks
References
Statistical mechanics of complex networks
SNAP: A General-Purpose Network Analysis and Graph-Mining Library
Family influences on the risk of daily smoking initiation.
Minimum connected dominating sets and maximal independent sets in unit disk graphs
Review: Connected dominating sets in wireless ad hoc and sensor networks - A comprehensive survey
Related Papers (5)
Frequently Asked Questions (15)
Q2. What are the future works mentioned in the paper "Positive influence dominating set generation in social networks" ?
For future work, the authors plan to further investigate the relationship between average degree and the PIDS and TPIDS size. In addition to this, their PIDS selection function g could be refined through further experimentation. Also, extensions of these algorithms to directed graphs that represent social networks like Twitter will be also be useful since these networks allow users to influence others without the reverse being true.
Q3. What is the simplest way to maximize f?
Then the node u that maximizes f will also be the one that touches the most unsatisfied nodes, since nodes that are already satisfied by P cannot contribute any more to the sum in f because of the min limitation, therefore the only way f will increase is if u touches unmet nodes.
Q4. What is the function that adds the number of unsatisfied neighbors?
The loop in line 3 will run at most n times (where n is the number of nodes in the graph), since the authors can only add as many nodes to their PIDS P as the authors have altogether, and each iteration of the loop adds a node from V to P .
Q5. What is the function that calculates the number of unsatisfied neighbors?
Their loop simply chooses the node with the maximum g value, adds that to P , and recomputes the g values for the remaining nodes that are not in P (denoted in Line 6 by the nodes in V − P ).
Q6. What is the way to generate scale-free networks?
The C + + SNAP library allowed us to easily generate scale-free networks using their implementation of the Barabasi-Albert model [13].
Q7. How can the authors test satisfaction under TPIDS?
To construct a TPIDS, the function s described in Section III-B, can also be adapted to test satisfaction under TPIDS, by simply removing the u ∈ P qualification, since in a TPIDS every node needs to have at least half of its neighbors in their TPIDS subset.
Q8. What is the function f() used to generate a TPIDS?
This function sums the number of neighbors that each node has in the TPIDSin-progress P (given by the function nP (u)), with the stipulation that each node can only contribute up to half its neighbors to the sum.
Q9. Why is the PIDS problem more accurate?
This is because the PIDS problem is a more accurate representation of spreading a positive message since nodes participating in spreading the message are already considered to be influenced and thus do not need half their neighbors to repeat the message back to them.
Q10. What is the way to use the algorithm to influence others?
extensions of these algorithms to directed graphs thatrepresent social networks like Twitter will be also be useful since these networks allow users to influence others without the reverse being true.
Q11. What is the first thing the authors do in the loop?
The first thing the authors do in the loop is to check if P is a PIDS or not, and only continue if the latter is true (this also takes place on line 3).
Q12. what is the algorithm for generating a PIDS?
Their algorithm which the authors call AltGreedy can be written as follows.1: P ← ∅ 2: compute g(u) value for all u ∈ V 3: while P is not a PIDS do 4: select u ∈ V − P to maximize g(u) 5: and set P ← P ∪ {u} 6: revise g(w) values for all w ∈ V − P 7: end while 8: return PTheir algorithm’s strategy for generating a PIDS is to be greedy and select the node that is connected to the most unsatisfied nodes, and add that node to the PIDS underconstruction represented by the set P .
Q13. What is the definition of a PIDS?
They collapsed neutral and negative into a single negative category, and then used the two remaining categories (positive and negative) as a starting point for their PIDS computation.
Q14. What is the topic of dominating sets?
While the topic of dominating sets has several years of research behind it in particular for backbone formation in Wireless Sensor Networks [7] [8] [9], the problem of Positive Influence Dominating Sets (PIDS) in social networks has been studied relatively recently.
Q15. How did the authors test the three algorithms?
In order to evaluate performance, all three algorithms - WangGreedy, RaeiGreedy and AltGreedy were tested by having them generate both a Positive Influence Dominating Set (PIDS) and a Total Positive Influence Dominating Set (TPIDS) on random scale-free graphs.