On feature relevance in image-based prediction models: an empirical study

Ender Konukoglu; Melanie Ganz; Koen Van Leemput; Mert R. Sabuncu

doi:10.1007/978-3-319-02267-3_22

On feature relevance in image-based prediction models: an empirical study

Ender Konukoglu, Melanie Ganz, Koen Van Leemput, Mert R. Sabuncu

2 Citations (Scopus)

Abstract

Determining disease-related variations of the anatomy and function is an important step in better understanding diseases and developing early diagnostic systems. In particular, image-based multivariate prediction models and the "relevant features" they produce are attracting attention from the community. In this article, we present an empirical study on the relevant features produced by two recently developed discriminative learning algorithms: neighborhood approximation forests (NAF) and the relevance voxel machine (RVoxM). Specifically, we examine whether the sets of features these methods produce are exhaustive; that is whether the features that are not marked as relevant carry disease-related information. We perform experiments on three different problems: image-based regression on a synthetic dataset for which the set of relevant features is known, regression of subject age as well as binary classification of Alzheimer's Disease (AD) from brain Magnetic Resonance Imaging (MRI) data. Our experiments demonstrate that aging-related and AD-related variations are widespread and the initial sets of relevant features discovered by the methods are not exhaustive. Our findings show that by knocking-out features and re-training models, a much larger set of disease-related features can be identified.

Original language	English
Title of host publication	Machine Learning in Medical Imaging : 4th International Workshop, MLMI 2013, Held in Conjunction with MICCAI 2013, Nagoya, Japan, September 22, 2013. Proceedings
Editors	Guorong Wu, Daoqiang Zhang, Dinggang Shen, Pingkun Yan, Kenji Suzuki, Fei Wang
Number of pages	8
Publisher	Springer Publishing Company
Publication date	2013
Pages	171-178
ISBN (Print)	978-3-319-02266-6
ISBN (Electronic)	978-3-319-02267-3
DOIs	https://doi.org/10.1007/978-3-319-02267-3_22
Publication status	Published - 2013
Externally published	Yes
Event	4th International Workshop on Machine Learning in Medical Imaging - Nagoya, Japan Duration: 22 Sept 2013 → 22 Sept 2013 Conference number: 4

Conference

Conference	4th International Workshop on Machine Learning in Medical Imaging
Number	4
Country/Territory	Japan
City	Nagoya
Period	22/09/2013 → 22/09/2013

Series	Lecture notes in computer science
Volume	8184
ISSN	0302-9743

Access to Document

10.1007/978-3-319-02267-3_22

Cite this

Konukoglu, E., Ganz, M., Van Leemput, K., & Sabuncu, M. R. (2013). On feature relevance in image-based prediction models: an empirical study. In G. Wu, D. Zhang, D. Shen, P. Yan, K. Suzuki, & F. Wang (Eds.), Machine Learning in Medical Imaging: 4th International Workshop, MLMI 2013, Held in Conjunction with MICCAI 2013, Nagoya, Japan, September 22, 2013. Proceedings (pp. 171-178). Springer Publishing Company. https://doi.org/10.1007/978-3-319-02267-3_22

On feature relevance in image-based prediction models: an empirical study. / Konukoglu, Ender; Ganz, Melanie; Van Leemput, Koen et al.
Machine Learning in Medical Imaging: 4th International Workshop, MLMI 2013, Held in Conjunction with MICCAI 2013, Nagoya, Japan, September 22, 2013. Proceedings. ed. / Guorong Wu; Daoqiang Zhang; Dinggang Shen; Pingkun Yan; Kenji Suzuki; Fei Wang. Springer Publishing Company, 2013. p. 171-178 (Lecture notes in computer science, Vol. 8184).

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

Konukoglu, E, Ganz, M, Van Leemput, K & Sabuncu, MR 2013, On feature relevance in image-based prediction models: an empirical study. in G Wu, D Zhang, D Shen, P Yan, K Suzuki & F Wang (eds), Machine Learning in Medical Imaging: 4th International Workshop, MLMI 2013, Held in Conjunction with MICCAI 2013, Nagoya, Japan, September 22, 2013. Proceedings. Springer Publishing Company, Lecture notes in computer science, vol. 8184, pp. 171-178, 4th International Workshop on Machine Learning in Medical Imaging, Nagoya, Japan, 22/09/2013. https://doi.org/10.1007/978-3-319-02267-3_22

Konukoglu E, Ganz M, Van Leemput K, Sabuncu MR. On feature relevance in image-based prediction models: an empirical study. In Wu G, Zhang D, Shen D, Yan P, Suzuki K, Wang F, editors, Machine Learning in Medical Imaging: 4th International Workshop, MLMI 2013, Held in Conjunction with MICCAI 2013, Nagoya, Japan, September 22, 2013. Proceedings. Springer Publishing Company. 2013. p. 171-178. (Lecture notes in computer science, Vol. 8184). doi: 10.1007/978-3-319-02267-3_22

Konukoglu, Ender ; Ganz, Melanie ; Van Leemput, Koen et al. / On feature relevance in image-based prediction models : an empirical study. Machine Learning in Medical Imaging: 4th International Workshop, MLMI 2013, Held in Conjunction with MICCAI 2013, Nagoya, Japan, September 22, 2013. Proceedings. editor / Guorong Wu ; Daoqiang Zhang ; Dinggang Shen ; Pingkun Yan ; Kenji Suzuki ; Fei Wang. Springer Publishing Company, 2013. pp. 171-178 (Lecture notes in computer science, Vol. 8184).

@inproceedings{543bd98a313e42fcbb2b37e677dc6251,

title = "On feature relevance in image-based prediction models: an empirical study",

abstract = "Determining disease-related variations of the anatomy and function is an important step in better understanding diseases and developing early diagnostic systems. In particular, image-based multivariate prediction models and the {"}relevant features{"} they produce are attracting attention from the community. In this article, we present an empirical study on the relevant features produced by two recently developed discriminative learning algorithms: neighborhood approximation forests (NAF) and the relevance voxel machine (RVoxM). Specifically, we examine whether the sets of features these methods produce are exhaustive; that is whether the features that are not marked as relevant carry disease-related information. We perform experiments on three different problems: image-based regression on a synthetic dataset for which the set of relevant features is known, regression of subject age as well as binary classification of Alzheimer's Disease (AD) from brain Magnetic Resonance Imaging (MRI) data. Our experiments demonstrate that aging-related and AD-related variations are widespread and the initial sets of relevant features discovered by the methods are not exhaustive. Our findings show that by knocking-out features and re-training models, a much larger set of disease-related features can be identified.",

author = "Ender Konukoglu and Melanie Ganz and {Van Leemput}, Koen and Sabuncu, {Mert R.}",

year = "2013",

doi = "10.1007/978-3-319-02267-3_22",

language = "English",

isbn = "978-3-319-02266-6 ",

series = "Lecture notes in computer science",

publisher = "Springer Publishing Company",

pages = "171--178",

editor = "Guorong Wu and Daoqiang Zhang and Dinggang Shen and Pingkun Yan and Kenji Suzuki and Fei Wang",

booktitle = "Machine Learning in Medical Imaging",

note = "4th International Workshop on Machine Learning in Medical Imaging, MLMI 2013 ; Conference date: 22-09-2013 Through 22-09-2013",

}

TY - GEN

T1 - On feature relevance in image-based prediction models

T2 - 4th International Workshop on Machine Learning in Medical Imaging

AU - Konukoglu, Ender

AU - Ganz, Melanie

AU - Van Leemput, Koen

AU - Sabuncu, Mert R.

N1 - Conference code: 4

PY - 2013

Y1 - 2013

N2 - Determining disease-related variations of the anatomy and function is an important step in better understanding diseases and developing early diagnostic systems. In particular, image-based multivariate prediction models and the "relevant features" they produce are attracting attention from the community. In this article, we present an empirical study on the relevant features produced by two recently developed discriminative learning algorithms: neighborhood approximation forests (NAF) and the relevance voxel machine (RVoxM). Specifically, we examine whether the sets of features these methods produce are exhaustive; that is whether the features that are not marked as relevant carry disease-related information. We perform experiments on three different problems: image-based regression on a synthetic dataset for which the set of relevant features is known, regression of subject age as well as binary classification of Alzheimer's Disease (AD) from brain Magnetic Resonance Imaging (MRI) data. Our experiments demonstrate that aging-related and AD-related variations are widespread and the initial sets of relevant features discovered by the methods are not exhaustive. Our findings show that by knocking-out features and re-training models, a much larger set of disease-related features can be identified.

AB - Determining disease-related variations of the anatomy and function is an important step in better understanding diseases and developing early diagnostic systems. In particular, image-based multivariate prediction models and the "relevant features" they produce are attracting attention from the community. In this article, we present an empirical study on the relevant features produced by two recently developed discriminative learning algorithms: neighborhood approximation forests (NAF) and the relevance voxel machine (RVoxM). Specifically, we examine whether the sets of features these methods produce are exhaustive; that is whether the features that are not marked as relevant carry disease-related information. We perform experiments on three different problems: image-based regression on a synthetic dataset for which the set of relevant features is known, regression of subject age as well as binary classification of Alzheimer's Disease (AD) from brain Magnetic Resonance Imaging (MRI) data. Our experiments demonstrate that aging-related and AD-related variations are widespread and the initial sets of relevant features discovered by the methods are not exhaustive. Our findings show that by knocking-out features and re-training models, a much larger set of disease-related features can be identified.

U2 - 10.1007/978-3-319-02267-3_22

DO - 10.1007/978-3-319-02267-3_22

M3 - Article in proceedings

SN - 978-3-319-02266-6

T3 - Lecture notes in computer science

SP - 171

EP - 178

BT - Machine Learning in Medical Imaging

A2 - Wu, Guorong

A2 - Zhang, Daoqiang

A2 - Shen, Dinggang

A2 - Yan, Pingkun

A2 - Suzuki, Kenji

A2 - Wang, Fei

PB - Springer Publishing Company

Y2 - 22 September 2013 through 22 September 2013

ER -

On feature relevance in image-based prediction models: an empirical study

Abstract

Conference

Access to Document

Fingerprint

Cite this