### Abstract

Preventing contagion in networks is an important problem in public health and other domains. Targeting nodes to immunize based on their network interactions has been shown to be far more effective at stemming infection spread than immunizing random subsets of nodes. However, the assumption that selected nodes can be rendered completely immune does not hold for infections for which there is no vaccination or effective treatment. Instead, one can confer fractional immunity to some nodes by allocating variable amounts of infection-prevention resource to them. We formulate the problem to distribute a fixed amount of resource across nodes in a network such that the infection rate is minimized, prove that it is NP-complete and derive a highly effective and efficient linear-time algorithm. We demonstrate the efficiency and accuracy of our algorithm compared to several other methods using simulation on realworld network datasets including US-MEDICARE and state-level interhospital patient transfer data. We find that concentrating resources at a small subset of nodes using our algorithm is up to 6 times more effective than distributing them uniformly (as is current practice) or using network-based heuristics. To the best of our knowledge, we are the first to formulate the problem, use truly nation-scale network data and propose effective algorithms.

Original language | English (US) |
---|---|

Title of host publication | Proceedings of the 2013 SIAM International Conference on Data Mining, SDM 2013 |

Publisher | Siam Society |

Pages | 659-667 |

Number of pages | 9 |

ISBN (Print) | 9781611972627 |

State | Published - 2013 |

Externally published | Yes |

Event | SIAM International Conference on Data Mining, SDM 2013 - Austin, United States Duration: May 2 2013 → May 4 2013 |

### Other

Other | SIAM International Conference on Data Mining, SDM 2013 |
---|---|

Country | United States |

City | Austin |

Period | 5/2/13 → 5/4/13 |

### Fingerprint

### ASJC Scopus subject areas

- Computer Science Applications
- Software

### Cite this

*Proceedings of the 2013 SIAM International Conference on Data Mining, SDM 2013*(pp. 659-667). Siam Society.

**Fractional immunization in networks.** / Prakash, B. Aditya; Adamic, Lada; Iwashyna, Theodore; Tong, Hanghang; Faloutsos, Christos.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

*Proceedings of the 2013 SIAM International Conference on Data Mining, SDM 2013.*Siam Society, pp. 659-667, SIAM International Conference on Data Mining, SDM 2013, Austin, United States, 5/2/13.

}

TY - GEN

T1 - Fractional immunization in networks

AU - Prakash, B. Aditya

AU - Adamic, Lada

AU - Iwashyna, Theodore

AU - Tong, Hanghang

AU - Faloutsos, Christos

PY - 2013

Y1 - 2013

N2 - Preventing contagion in networks is an important problem in public health and other domains. Targeting nodes to immunize based on their network interactions has been shown to be far more effective at stemming infection spread than immunizing random subsets of nodes. However, the assumption that selected nodes can be rendered completely immune does not hold for infections for which there is no vaccination or effective treatment. Instead, one can confer fractional immunity to some nodes by allocating variable amounts of infection-prevention resource to them. We formulate the problem to distribute a fixed amount of resource across nodes in a network such that the infection rate is minimized, prove that it is NP-complete and derive a highly effective and efficient linear-time algorithm. We demonstrate the efficiency and accuracy of our algorithm compared to several other methods using simulation on realworld network datasets including US-MEDICARE and state-level interhospital patient transfer data. We find that concentrating resources at a small subset of nodes using our algorithm is up to 6 times more effective than distributing them uniformly (as is current practice) or using network-based heuristics. To the best of our knowledge, we are the first to formulate the problem, use truly nation-scale network data and propose effective algorithms.

AB - Preventing contagion in networks is an important problem in public health and other domains. Targeting nodes to immunize based on their network interactions has been shown to be far more effective at stemming infection spread than immunizing random subsets of nodes. However, the assumption that selected nodes can be rendered completely immune does not hold for infections for which there is no vaccination or effective treatment. Instead, one can confer fractional immunity to some nodes by allocating variable amounts of infection-prevention resource to them. We formulate the problem to distribute a fixed amount of resource across nodes in a network such that the infection rate is minimized, prove that it is NP-complete and derive a highly effective and efficient linear-time algorithm. We demonstrate the efficiency and accuracy of our algorithm compared to several other methods using simulation on realworld network datasets including US-MEDICARE and state-level interhospital patient transfer data. We find that concentrating resources at a small subset of nodes using our algorithm is up to 6 times more effective than distributing them uniformly (as is current practice) or using network-based heuristics. To the best of our knowledge, we are the first to formulate the problem, use truly nation-scale network data and propose effective algorithms.

UR - http://www.scopus.com/inward/record.url?scp=84934293884&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84934293884&partnerID=8YFLogxK

M3 - Conference contribution

SN - 9781611972627

SP - 659

EP - 667

BT - Proceedings of the 2013 SIAM International Conference on Data Mining, SDM 2013

PB - Siam Society

ER -