Seeding with Costly Network Information
Name
1905.04325.pdf
Description
Accepted version
Size
2.01 MB
Format
Adobe PDF
Checksum (MD5)
ea77bccc57814afa48f492ff254f61da
Author(s)
Eckles, Dean
Esfandiari, Hossein
Mossel, Elchanan
Rahimian, M Amin
Date Issued
2022
Journal
Operations Research
Publisher
Institute for Operations Research and the Management Sciences (INFORMS)
Citation
Eckles, Dean, Esfandiari, Hossein, Mossel, Elchanan and Rahimian, M Amin. 2022. "Seeding with Costly Network Information." Operations Research, 70 (4).
Version
Author's final manuscript
Abstract
In the presence of contagion, decision makers strategize about where in a network to intervene (e.g., seeding a new product). A large literature has developed methods for approximately optimizing the choice of k seeds to cause the largest cascade of, for example, product adoption. However, it is often impractical to measure an entire social network. In “Seeding with Costly Network Information,” Eckles, Esfandiari, Mossel, and Rahimian develop and analyze algorithms for making a bounded number of queries of a social network and then selecting k seeds. They prove hardness results for this problem and provide almost tight approximation guarantees for their proposed algorithms under widely used models of contagion. One proposed algorithm is practical for both querying online social networks and structuring in-person surveys. This framework further allows reasoning about tradeoffs between spending budget on collecting more network data versus increasing the number of seeds.
MIT Department
Massachusetts Institute of Technology. Department of Mathematics
Sloan School of Management
Terms of Use
Creative Commons Attribution-Noncommercial-Share Alike
Persistent DSpace Link
DOI of Published Version
https://doi.org/10.1287/OPRE.2022.2290
