MultiCost: Multi-stage cost-sensitive classification of Alzheimer's disease

Daoqiang Zhang, Dinggang Shen

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

5 Citations (Scopus)

Abstract

Most traditional classification methods for Alzheimer's disease (AD) aim to obtain a high accuracy, or equivalently a low classification error rate, which implicitly assumes that the losses of all misclassifications are the same. However, in practical AD diagnosis, the losses of misclassifying healthy subjects and AD patients are usually very different. For example, it may be troublesome if a healthy subject is misclassified as AD, but it could result in a more serious consequence if an AD patient is misclassified as healthy subject. In this paper, we propose a multi-stage cost-sensitive approach for AD classification via multimodal imaging data and CSF biomarkers. Our approach contains three key components: (1) a cost-sensitive feature selection which can select more AD-related brain regions by using different costs for different misclassifications in the feature selection stage, (2) a multimodal data fusion which effectively fuses data from MRI, PET and CSF biomarkers based on multiple kernels combination, and (3) a cost-sensitive classifier construction which further reduces the overall misclassification loss through a threshold-moving strategy. Experimental results on ADNI dataset show that the proposed approach can significantly reduce the cost of misclassification and simultaneously improve the sensitivity, under the same or even higher classification accuracy compared with conventional methods.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages344-351
Number of pages8
Volume7009 LNCS
DOIs
Publication statusPublished - 2011 Oct 17
Externally publishedYes
Event2nd International Workshop on Machine Learning in Medical Imaging, MLMI 2011, in Conjunction with the 14th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2011 - Toronto, ON, Canada
Duration: 2011 Sep 182011 Sep 18

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume7009 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other2nd International Workshop on Machine Learning in Medical Imaging, MLMI 2011, in Conjunction with the 14th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2011
CountryCanada
CityToronto, ON
Period11/9/1811/9/18

Fingerprint

Alzheimer's Disease
Misclassification
Costs
Biomarkers
Feature Selection
Feature extraction
Data Fusion
Data fusion
Electric fuses
Magnetic resonance imaging
Error Rate
Brain
High Accuracy
Classifiers
Classifier
Imaging
kernel
Imaging techniques
Experimental Results

Keywords

  • Alzheimer's disease (AD)
  • Cost-sensitive classification
  • cost-sensitive feature selection
  • multi-modality
  • MultiCost

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Zhang, D., & Shen, D. (2011). MultiCost: Multi-stage cost-sensitive classification of Alzheimer's disease. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 7009 LNCS, pp. 344-351). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7009 LNCS). https://doi.org/10.1007/978-3-642-24319-6_42

MultiCost : Multi-stage cost-sensitive classification of Alzheimer's disease. / Zhang, Daoqiang; Shen, Dinggang.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 7009 LNCS 2011. p. 344-351 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7009 LNCS).

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

Zhang, D & Shen, D 2011, MultiCost: Multi-stage cost-sensitive classification of Alzheimer's disease. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 7009 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 7009 LNCS, pp. 344-351, 2nd International Workshop on Machine Learning in Medical Imaging, MLMI 2011, in Conjunction with the 14th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2011, Toronto, ON, Canada, 11/9/18. https://doi.org/10.1007/978-3-642-24319-6_42
Zhang D, Shen D. MultiCost: Multi-stage cost-sensitive classification of Alzheimer's disease. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 7009 LNCS. 2011. p. 344-351. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-642-24319-6_42
Zhang, Daoqiang ; Shen, Dinggang. / MultiCost : Multi-stage cost-sensitive classification of Alzheimer's disease. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 7009 LNCS 2011. pp. 344-351 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{426b8dced5c740a18cc414e6f4d1f19e,
title = "MultiCost: Multi-stage cost-sensitive classification of Alzheimer's disease",
abstract = "Most traditional classification methods for Alzheimer's disease (AD) aim to obtain a high accuracy, or equivalently a low classification error rate, which implicitly assumes that the losses of all misclassifications are the same. However, in practical AD diagnosis, the losses of misclassifying healthy subjects and AD patients are usually very different. For example, it may be troublesome if a healthy subject is misclassified as AD, but it could result in a more serious consequence if an AD patient is misclassified as healthy subject. In this paper, we propose a multi-stage cost-sensitive approach for AD classification via multimodal imaging data and CSF biomarkers. Our approach contains three key components: (1) a cost-sensitive feature selection which can select more AD-related brain regions by using different costs for different misclassifications in the feature selection stage, (2) a multimodal data fusion which effectively fuses data from MRI, PET and CSF biomarkers based on multiple kernels combination, and (3) a cost-sensitive classifier construction which further reduces the overall misclassification loss through a threshold-moving strategy. Experimental results on ADNI dataset show that the proposed approach can significantly reduce the cost of misclassification and simultaneously improve the sensitivity, under the same or even higher classification accuracy compared with conventional methods.",
keywords = "Alzheimer's disease (AD), Cost-sensitive classification, cost-sensitive feature selection, multi-modality, MultiCost",
author = "Daoqiang Zhang and Dinggang Shen",
year = "2011",
month = "10",
day = "17",
doi = "10.1007/978-3-642-24319-6_42",
language = "English",
isbn = "9783642243189",
volume = "7009 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "344--351",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - MultiCost

T2 - Multi-stage cost-sensitive classification of Alzheimer's disease

AU - Zhang, Daoqiang

AU - Shen, Dinggang

PY - 2011/10/17

Y1 - 2011/10/17

N2 - Most traditional classification methods for Alzheimer's disease (AD) aim to obtain a high accuracy, or equivalently a low classification error rate, which implicitly assumes that the losses of all misclassifications are the same. However, in practical AD diagnosis, the losses of misclassifying healthy subjects and AD patients are usually very different. For example, it may be troublesome if a healthy subject is misclassified as AD, but it could result in a more serious consequence if an AD patient is misclassified as healthy subject. In this paper, we propose a multi-stage cost-sensitive approach for AD classification via multimodal imaging data and CSF biomarkers. Our approach contains three key components: (1) a cost-sensitive feature selection which can select more AD-related brain regions by using different costs for different misclassifications in the feature selection stage, (2) a multimodal data fusion which effectively fuses data from MRI, PET and CSF biomarkers based on multiple kernels combination, and (3) a cost-sensitive classifier construction which further reduces the overall misclassification loss through a threshold-moving strategy. Experimental results on ADNI dataset show that the proposed approach can significantly reduce the cost of misclassification and simultaneously improve the sensitivity, under the same or even higher classification accuracy compared with conventional methods.

AB - Most traditional classification methods for Alzheimer's disease (AD) aim to obtain a high accuracy, or equivalently a low classification error rate, which implicitly assumes that the losses of all misclassifications are the same. However, in practical AD diagnosis, the losses of misclassifying healthy subjects and AD patients are usually very different. For example, it may be troublesome if a healthy subject is misclassified as AD, but it could result in a more serious consequence if an AD patient is misclassified as healthy subject. In this paper, we propose a multi-stage cost-sensitive approach for AD classification via multimodal imaging data and CSF biomarkers. Our approach contains three key components: (1) a cost-sensitive feature selection which can select more AD-related brain regions by using different costs for different misclassifications in the feature selection stage, (2) a multimodal data fusion which effectively fuses data from MRI, PET and CSF biomarkers based on multiple kernels combination, and (3) a cost-sensitive classifier construction which further reduces the overall misclassification loss through a threshold-moving strategy. Experimental results on ADNI dataset show that the proposed approach can significantly reduce the cost of misclassification and simultaneously improve the sensitivity, under the same or even higher classification accuracy compared with conventional methods.

KW - Alzheimer's disease (AD)

KW - Cost-sensitive classification

KW - cost-sensitive feature selection

KW - multi-modality

KW - MultiCost

UR - http://www.scopus.com/inward/record.url?scp=80053959264&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=80053959264&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-24319-6_42

DO - 10.1007/978-3-642-24319-6_42

M3 - Conference contribution

AN - SCOPUS:80053959264

SN - 9783642243189

VL - 7009 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 344

EP - 351

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

ER -