Towards Confidence Interval Estimation in Truth Discovery

Houping Xiao, Jing Gao, Qi Li, Fenglong Ma, Lu Su, Yunlong Feng, Aidong Zhang

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

The demand for automatic extraction of true information (i.e., truths) from conflicting multi-source data has soared recently. A variety of truth discovery methods have witnessed great successes via jointly estimating source reliability and truths. All existing truth discovery methods focus on providing a point estimator for each object's truth, but in many real-world applications, confidence interval estimation of truths is more desirable, since confidence interval contains richer information. To address this challenge, in this paper, we propose a novel truth discovery method (ETCIBoot) to construct confidence interval estimates as well as identify truths, where the bootstrapping techniques are nicely integrated into the truth discovery procedure. Due to the properties of bootstrapping, the estimators obtained by ETCIBoot are more accurate and robust compared with the state-of-the-art truth discovery approaches. The proposed framework is further adapted to deal with large-scale truth discovery task in distributed paradigm. Theoretically, we prove the asymptotical consistency of the confidence interval obtained by ETCIBoot. Experimentally, we demonstrate that ETCIBoot is not only effective in constructing confidence intervals but also able to obtain better truth estimates.

Original languageEnglish (US)
Article number8359426
Pages (from-to)575-588
Number of pages14
JournalIEEE Transactions on Knowledge and Data Engineering
Volume31
Issue number3
DOIs
StatePublished - Mar 1 2019

All Science Journal Classification (ASJC) codes

  • Information Systems
  • Computer Science Applications
  • Computational Theory and Mathematics

Cite this

Xiao, Houping ; Gao, Jing ; Li, Qi ; Ma, Fenglong ; Su, Lu ; Feng, Yunlong ; Zhang, Aidong. / Towards Confidence Interval Estimation in Truth Discovery. In: IEEE Transactions on Knowledge and Data Engineering. 2019 ; Vol. 31, No. 3. pp. 575-588.
@article{a62ac3976e2948ad9584322291400e32,
title = "Towards Confidence Interval Estimation in Truth Discovery",
abstract = "The demand for automatic extraction of true information (i.e., truths) from conflicting multi-source data has soared recently. A variety of truth discovery methods have witnessed great successes via jointly estimating source reliability and truths. All existing truth discovery methods focus on providing a point estimator for each object's truth, but in many real-world applications, confidence interval estimation of truths is more desirable, since confidence interval contains richer information. To address this challenge, in this paper, we propose a novel truth discovery method (ETCIBoot) to construct confidence interval estimates as well as identify truths, where the bootstrapping techniques are nicely integrated into the truth discovery procedure. Due to the properties of bootstrapping, the estimators obtained by ETCIBoot are more accurate and robust compared with the state-of-the-art truth discovery approaches. The proposed framework is further adapted to deal with large-scale truth discovery task in distributed paradigm. Theoretically, we prove the asymptotical consistency of the confidence interval obtained by ETCIBoot. Experimentally, we demonstrate that ETCIBoot is not only effective in constructing confidence intervals but also able to obtain better truth estimates.",
author = "Houping Xiao and Jing Gao and Qi Li and Fenglong Ma and Lu Su and Yunlong Feng and Aidong Zhang",
year = "2019",
month = "3",
day = "1",
doi = "10.1109/TKDE.2018.2837026",
language = "English (US)",
volume = "31",
pages = "575--588",
journal = "IEEE Transactions on Knowledge and Data Engineering",
issn = "1041-4347",
publisher = "IEEE Computer Society",
number = "3",

}

Towards Confidence Interval Estimation in Truth Discovery. / Xiao, Houping; Gao, Jing; Li, Qi; Ma, Fenglong; Su, Lu; Feng, Yunlong; Zhang, Aidong.

In: IEEE Transactions on Knowledge and Data Engineering, Vol. 31, No. 3, 8359426, 01.03.2019, p. 575-588.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Towards Confidence Interval Estimation in Truth Discovery

AU - Xiao, Houping

AU - Gao, Jing

AU - Li, Qi

AU - Ma, Fenglong

AU - Su, Lu

AU - Feng, Yunlong

AU - Zhang, Aidong

PY - 2019/3/1

Y1 - 2019/3/1

N2 - The demand for automatic extraction of true information (i.e., truths) from conflicting multi-source data has soared recently. A variety of truth discovery methods have witnessed great successes via jointly estimating source reliability and truths. All existing truth discovery methods focus on providing a point estimator for each object's truth, but in many real-world applications, confidence interval estimation of truths is more desirable, since confidence interval contains richer information. To address this challenge, in this paper, we propose a novel truth discovery method (ETCIBoot) to construct confidence interval estimates as well as identify truths, where the bootstrapping techniques are nicely integrated into the truth discovery procedure. Due to the properties of bootstrapping, the estimators obtained by ETCIBoot are more accurate and robust compared with the state-of-the-art truth discovery approaches. The proposed framework is further adapted to deal with large-scale truth discovery task in distributed paradigm. Theoretically, we prove the asymptotical consistency of the confidence interval obtained by ETCIBoot. Experimentally, we demonstrate that ETCIBoot is not only effective in constructing confidence intervals but also able to obtain better truth estimates.

AB - The demand for automatic extraction of true information (i.e., truths) from conflicting multi-source data has soared recently. A variety of truth discovery methods have witnessed great successes via jointly estimating source reliability and truths. All existing truth discovery methods focus on providing a point estimator for each object's truth, but in many real-world applications, confidence interval estimation of truths is more desirable, since confidence interval contains richer information. To address this challenge, in this paper, we propose a novel truth discovery method (ETCIBoot) to construct confidence interval estimates as well as identify truths, where the bootstrapping techniques are nicely integrated into the truth discovery procedure. Due to the properties of bootstrapping, the estimators obtained by ETCIBoot are more accurate and robust compared with the state-of-the-art truth discovery approaches. The proposed framework is further adapted to deal with large-scale truth discovery task in distributed paradigm. Theoretically, we prove the asymptotical consistency of the confidence interval obtained by ETCIBoot. Experimentally, we demonstrate that ETCIBoot is not only effective in constructing confidence intervals but also able to obtain better truth estimates.

UR - http://www.scopus.com/inward/record.url?scp=85047019943&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85047019943&partnerID=8YFLogxK

U2 - 10.1109/TKDE.2018.2837026

DO - 10.1109/TKDE.2018.2837026

M3 - Article

AN - SCOPUS:85047019943

VL - 31

SP - 575

EP - 588

JO - IEEE Transactions on Knowledge and Data Engineering

JF - IEEE Transactions on Knowledge and Data Engineering

SN - 1041-4347

IS - 3

M1 - 8359426

ER -