schliessen

Filtern

 

Bibliotheken

Network-Constrained Group Lasso for High-Dimensional Multinomial Classification with Application to Cancer Subtype Prediction

Classic multinomial logit model, commonly used in multiclass regression problem, is restricted to few predictors and does not take into account the relationship among variables. It has limited use for genomic data, where the number of genomic features far exceeds the sample size. Genomic features su... Full description

Journal Title: Cancer Informatics January 2014, Vol.13s6
Main Author: Tian, Xinyu
Other Authors: Wang, Xuefeng , Chen, Jun
Format: Electronic Article Electronic Article
Language: English
Subjects:
ID: E-ISSN: 1176-9351 ; DOI: 10.4137/CIN.S17686
Link: https://journals.sagepub.com/doi/full/10.4137/CIN.S17686
Zum Text:
SendSend as email Add to Book BagAdd to Book Bag
Staff View
recordid: sage_s10_4137_CIN_S17686
title: Network-Constrained Group Lasso for High-Dimensional Multinomial Classification with Application to Cancer Subtype Prediction
format: Article
creator:
  • Tian, Xinyu
  • Wang, Xuefeng
  • Chen, Jun
subjects:
  • Cancer Subtype Prediction
  • Multinomial Logit Model
  • Group Lasso
  • Network-Constraint
  • Proximal Gradient Algorithm
  • Medicine
ispartof: Cancer Informatics, January 2014, Vol.13s6
description: Classic multinomial logit model, commonly used in multiclass regression problem, is restricted to few predictors and does not take into account the relationship among variables. It has limited use for genomic data, where the number of genomic features far exceeds the sample size. Genomic features such as gene expressions are usually related by an underlying biological network. Efficient use of the network information is important to improve classification performance as well as the biological interpretability. We proposed a multinomial logit model that is capable of addressing both the high dimensionality of predictors and the underlying network information. Group lasso was used to induce model sparsity, and a network-constraint was imposed to induce the smoothness of the coefficients with respect to the underlying network structure. To deal with the non-smoothness of the objective function in optimization, we developed a proximal gradient algorithm for efficient computation. The proposed model was compared to models with no prior structure information in both simulations and a problem of cancer subtype prediction with real TCGA (the cancer genome atlas) gene expression data. The network-constrained mode outperformed the traditional ones in both cases.
language: eng
source:
identifier: E-ISSN: 1176-9351 ; DOI: 10.4137/CIN.S17686
fulltext: fulltext_linktorsrc
issn:
  • 1176-9351
  • 11769351
url: Link


@attributes
ID656132737
RANK0.07
NO1
SEARCH_ENGINEprimo_central_multiple_fe
SEARCH_ENGINE_TYPEPrimo Central Search Engine
LOCALfalse
PrimoNMBib
record
control
sourcerecordid10_4137_CIN_S17686
sourceidsage_s
recordidTN_sage_s10_4137_CIN_S17686
sourcesystemOther
dbid
0.2J
1.2M
2.2N
331U
431X
531Z
654M
7AACTG
8AAMPI
9AATBZ
10ABJUO
11ABQXT
12ACGEY
13ACGZU
14AECGH
15AEDTQ
16AFRWT
17AGWFA
18AMCVQ
19B8R
20B8Z
21B94
22DB~
23DF0
24DO-
25DV7
26J8X
27K.F
28Q7L
29Q81
30Q83
31SFC
32SFK
33SFT
34SGR
35SGV
36SGZ
37SHG
38SNB
39SPP
40SPV
41STM
pqid1652433396
display
typearticle
titleNetwork-Constrained Group Lasso for High-Dimensional Multinomial Classification with Application to Cancer Subtype Prediction
creatorTian, Xinyu ; Wang, Xuefeng ; Chen, Jun
ispartofCancer Informatics, January 2014, Vol.13s6
identifierE-ISSN: 1176-9351 ; DOI: 10.4137/CIN.S17686
subjectCancer Subtype Prediction ; Multinomial Logit Model ; Group Lasso ; Network-Constraint ; Proximal Gradient Algorithm ; Medicine
descriptionClassic multinomial logit model, commonly used in multiclass regression problem, is restricted to few predictors and does not take into account the relationship among variables. It has limited use for genomic data, where the number of genomic features far exceeds the sample size. Genomic features such as gene expressions are usually related by an underlying biological network. Efficient use of the network information is important to improve classification performance as well as the biological interpretability. We proposed a multinomial logit model that is capable of addressing both the high dimensionality of predictors and the underlying network information. Group lasso was used to induce model sparsity, and a network-constraint was imposed to induce the smoothness of the coefficients with respect to the underlying network structure. To deal with the non-smoothness of the objective function in optimization, we developed a proximal gradient algorithm for efficient computation. The proposed model was compared to models with no prior structure information in both simulations and a problem of cancer subtype prediction with real TCGA (the cancer genome atlas) gene expression data. The network-constrained mode outperformed the traditional ones in both cases.
languageeng
oafree_for_read
source
version7
lds50peer_reviewed
links
openurl$$Topenurl_article
openurlfulltext$$Topenurlfull_article
backlink$$Uhttps://journals.sagepub.com/doi/full/10.4137/CIN.S17686$$EView_record_in_Sage_(Access_to_full_text_may_be_restricted)
linktorsrc$$Uhttps://journals.sagepub.com/doi/full/10.4137/CIN.S17686$$EView_full_text_in_Sage
search
creatorcontrib
0Tian, Xinyu
1Wang, Xuefeng
2Chen, Jun
titleNetwork-Constrained Group Lasso for High-Dimensional Multinomial Classification with Application to Cancer Subtype Prediction
description

Classic multinomial logit model, commonly used in multiclass regression problem, is restricted to few predictors and does not take into account the relationship among variables. It has limited use for genomic data, where the number of genomic features far exceeds the sample size. Genomic features such as gene expressions are usually related by an underlying biological network. Efficient use of the network information is important to improve classification performance as well as the biological interpretability. We proposed a multinomial logit model that is capable of addressing both the high dimensionality of predictors and the underlying network information. Group lasso was used to induce model sparsity, and a network-constraint was imposed to induce the smoothness of the coefficients with respect to the underlying network structure. To deal with the non-smoothness of the objective function in optimization, we developed a proximal gradient algorithm for efficient computation. The proposed model was compared to models with no prior structure information in both simulations and a problem of cancer subtype prediction with real TCGA (the cancer genome atlas) gene expression data. The network-constrained mode outperformed the traditional ones in both cases.

subject
0Cancer Subtype Prediction
1Multinomial Logit Model
2Group Lasso
3Network-Constraint
4Proximal Gradient Algorithm
5Medicine
general
0English
1SAGE Publications
210.4137/CIN.S17686
3Sage Journals (Sage Publications)
4SAGE Health Sciences (Sage Publications)
5SAGE STM (Sage Publications)
6SAGE Open Access Journals (Sage Publications)
7SAGE Communication and Media Studies (Sage Publications)
8SAGE Journals (Sage Publications)
sourceidsage_s
recordidsage_s10_4137_CIN_S17686
issn
01176-9351
111769351
rsrctypearticle
creationdate2014
addtitleCancer Informatics
searchscope
0sage_full
1sage131
2sage214
3sage252
4sage128
5sage210
6sage249
7sage32
8sage254
9sage219
10sage255
11sage220
12sage221
13sage216
14sage134
15sage135
16sage133
17sage_oa
18sage256
19sage136
20sage129
21sage211
22sage250
23sage215
24sage132
25sage253
26sage42
27sage40
28sage130
29sage212
30sage251
31sage11
32sage208
33sage206
34sage127
35sage209
36sage248
37sage35
38sage126
39sage205
40sage33
41sage247
scope
0sage_full
1sage131
2sage214
3sage252
4sage128
5sage210
6sage249
7sage32
8sage254
9sage219
10sage255
11sage220
12sage221
13sage216
14sage134
15sage135
16sage133
17sage_oa
18sage256
19sage136
20sage129
21sage211
22sage250
23sage215
24sage132
25sage253
26sage42
27sage40
28sage130
29sage212
30sage251
31sage11
32sage208
33sage206
34sage127
35sage209
36sage248
37sage35
38sage126
39sage205
40sage33
41sage247
lsr44$$EView_record_in_Sage_(Access_to_full_text_may_be_restricted)
lsr45$$EView_full_text_in_Sage
tmp01
0SAGE Health Sciences (Sage Publications)
1Sage Journals (Sage Publications)
2SAGE STM (Sage Publications)
3SAGE Open Access Journals (Sage Publications)
4SAGE Communication and Media Studies (Sage Publications)
5SAGE Journals (Sage Publications)
tmp02
0.2J
1.2M
2.2N
331U
431X
531Z
654M
7AACTG
8AAMPI
9AATBZ
10ABJUO
11ABQXT
12ACGEY
13ACGZU
14AECGH
15AEDTQ
16AFRWT
17AGWFA
18AMCVQ
19B8R
20B8Z
21B94
22DB~
23DF0
24DO-
25DV7
26J8X
27K.F
28Q7L
29Q81
30Q83
31SFC
32SFK
33SFT
34SGR
35SGV
36SGZ
37SHG
38SNB
39SPP
40SPV
41STM
startdate20140101
enddate20140131
lsr40Cancer Informatics, January 2014, Vol.13s6
doi10.4137/CIN.S17686
citationvol 13s6
lsr30VSR-Enriched:[issue, pqid, pages]
sort
titleNetwork-Constrained Group Lasso for High-Dimensional Multinomial Classification with Application to Cancer Subtype Prediction
authorTian, Xinyu ; Wang, Xuefeng ; Chen, Jun
creationdate20140100
lso0120140100
facets
frbrgroupid8012999028673479348
frbrtype5
newrecords20171017
languageeng
topic
0Cancer Subtype Prediction
1Multinomial Logit Model
2Group Lasso
3Network-Constraint
4Proximal Gradient Algorithm
5Medicine
collectionSage Journals (Sage Publications)
prefilterarticles
rsrctypearticles
creatorcontrib
0Tian, Xinyu
1Wang, Xuefeng
2Chen, Jun
jtitleCancer Informatics
creationdate2014
toplevelpeer_reviewed
delivery
delcategoryRemote Search Resource
fulltextfulltext_linktorsrc
addata
aulast
0Tian
1Wang
2Chen
aufirst
0Xinyu
1Xuefeng
2Jun
auinitX
auinit1X
au
0Tian, Xinyu
1Wang, Xuefeng
2Chen, Jun
atitleNetwork-Constrained Group Lasso for High-Dimensional Multinomial Classification with Application to Cancer Subtype Prediction
jtitleCancer Informatics
risdate201401
volume13s6
eissn1176-9351
formatjournal
genrearticle
ristypeJOUR
abstract

Classic multinomial logit model, commonly used in multiclass regression problem, is restricted to few predictors and does not take into account the relationship among variables. It has limited use for genomic data, where the number of genomic features far exceeds the sample size. Genomic features such as gene expressions are usually related by an underlying biological network. Efficient use of the network information is important to improve classification performance as well as the biological interpretability. We proposed a multinomial logit model that is capable of addressing both the high dimensionality of predictors and the underlying network information. Group lasso was used to induce model sparsity, and a network-constraint was imposed to induce the smoothness of the coefficients with respect to the underlying network structure. To deal with the non-smoothness of the objective function in optimization, we developed a proximal gradient algorithm for efficient computation. The proposed model was compared to models with no prior structure information in both simulations and a problem of cancer subtype prediction with real TCGA (the cancer genome atlas) gene expression data. The network-constrained mode outperformed the traditional ones in both cases.

copLondon, England
pubSAGE Publications
doi10.4137/CIN.S17686
oafree_for_read
issueSuppl 6
pages25-33
date2014-01