Mining causal outliers using gaussian Bayesian networks

Sakshi Babbar, Sanjay Chawla

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

Outliers are often identified as data points which are ''rare'', ''isolated'', or far away from their nearest neighbours. In this paper we demonstrate that meaningful outliers, i.e., outliers which perhaps encode important or new information are those which violate causal relationships. We first build a Bayesian network which encode causal relationships between attributes and then identify those points as outliers which violate these causal relationships. Experiments on several data sets confirm that the outliers identified in this fashion are in some sense ''genuine'' as they reveal new information about the underlying data generating process.

Original languageEnglish
Title of host publicationProceedings - International Conference on Tools with Artificial Intelligence, ICTAI
Pages97-104
Number of pages8
Volume1
DOIs
Publication statusPublished - 2012
Externally publishedYes
Event2012 IEEE 24th International Conference on Tools with Artificial Intelligence, ICTAI 2012 - Athens
Duration: 7 Nov 20129 Nov 2012

Other

Other2012 IEEE 24th International Conference on Tools with Artificial Intelligence, ICTAI 2012
CityAthens
Period7/11/129/11/12

Fingerprint

Bayesian networks
Experiments

Keywords

  • Bayesian networks
  • Causality and Outliers

ASJC Scopus subject areas

  • Software
  • Artificial Intelligence
  • Computer Science Applications

Cite this

Babbar, S., & Chawla, S. (2012). Mining causal outliers using gaussian Bayesian networks. In Proceedings - International Conference on Tools with Artificial Intelligence, ICTAI (Vol. 1, pp. 97-104). [6495034] https://doi.org/10.1109/ICTAI.2012.22

Mining causal outliers using gaussian Bayesian networks. / Babbar, Sakshi; Chawla, Sanjay.

Proceedings - International Conference on Tools with Artificial Intelligence, ICTAI. Vol. 1 2012. p. 97-104 6495034.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Babbar, S & Chawla, S 2012, Mining causal outliers using gaussian Bayesian networks. in Proceedings - International Conference on Tools with Artificial Intelligence, ICTAI. vol. 1, 6495034, pp. 97-104, 2012 IEEE 24th International Conference on Tools with Artificial Intelligence, ICTAI 2012, Athens, 7/11/12. https://doi.org/10.1109/ICTAI.2012.22
Babbar S, Chawla S. Mining causal outliers using gaussian Bayesian networks. In Proceedings - International Conference on Tools with Artificial Intelligence, ICTAI. Vol. 1. 2012. p. 97-104. 6495034 https://doi.org/10.1109/ICTAI.2012.22
Babbar, Sakshi ; Chawla, Sanjay. / Mining causal outliers using gaussian Bayesian networks. Proceedings - International Conference on Tools with Artificial Intelligence, ICTAI. Vol. 1 2012. pp. 97-104
@inproceedings{b2f29464d69a4b2d9c69894708332a27,
title = "Mining causal outliers using gaussian Bayesian networks",
abstract = "Outliers are often identified as data points which are ''rare'', ''isolated'', or far away from their nearest neighbours. In this paper we demonstrate that meaningful outliers, i.e., outliers which perhaps encode important or new information are those which violate causal relationships. We first build a Bayesian network which encode causal relationships between attributes and then identify those points as outliers which violate these causal relationships. Experiments on several data sets confirm that the outliers identified in this fashion are in some sense ''genuine'' as they reveal new information about the underlying data generating process.",
keywords = "Bayesian networks, Causality and Outliers",
author = "Sakshi Babbar and Sanjay Chawla",
year = "2012",
doi = "10.1109/ICTAI.2012.22",
language = "English",
isbn = "9780769549156",
volume = "1",
pages = "97--104",
booktitle = "Proceedings - International Conference on Tools with Artificial Intelligence, ICTAI",

}

TY - GEN

T1 - Mining causal outliers using gaussian Bayesian networks

AU - Babbar, Sakshi

AU - Chawla, Sanjay

PY - 2012

Y1 - 2012

N2 - Outliers are often identified as data points which are ''rare'', ''isolated'', or far away from their nearest neighbours. In this paper we demonstrate that meaningful outliers, i.e., outliers which perhaps encode important or new information are those which violate causal relationships. We first build a Bayesian network which encode causal relationships between attributes and then identify those points as outliers which violate these causal relationships. Experiments on several data sets confirm that the outliers identified in this fashion are in some sense ''genuine'' as they reveal new information about the underlying data generating process.

AB - Outliers are often identified as data points which are ''rare'', ''isolated'', or far away from their nearest neighbours. In this paper we demonstrate that meaningful outliers, i.e., outliers which perhaps encode important or new information are those which violate causal relationships. We first build a Bayesian network which encode causal relationships between attributes and then identify those points as outliers which violate these causal relationships. Experiments on several data sets confirm that the outliers identified in this fashion are in some sense ''genuine'' as they reveal new information about the underlying data generating process.

KW - Bayesian networks

KW - Causality and Outliers

UR - http://www.scopus.com/inward/record.url?scp=84876931267&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84876931267&partnerID=8YFLogxK

U2 - 10.1109/ICTAI.2012.22

DO - 10.1109/ICTAI.2012.22

M3 - Conference contribution

AN - SCOPUS:84876931267

SN - 9780769549156

VL - 1

SP - 97

EP - 104

BT - Proceedings - International Conference on Tools with Artificial Intelligence, ICTAI

ER -