Deep neural networks for choice analysis: A statistical learning theory perspective
•Used statistical learning theory to evaluate DNNs in choice analysis.•Operationalized DNN interpretability by using the choice probability functions.•Provided a tight upper bound on the estimation error of DNNs.•Conducted experiments to identify when DNNs outperform classical models.•DNNs can be mo...
Gespeichert in:
| Veröffentlicht in: | Transportation research. Part B: methodological Jg. 148; S. 60 - 81 |
|---|---|
| Hauptverfasser: | , , , |
| Format: | Journal Article |
| Sprache: | Englisch |
| Veröffentlicht: |
Oxford
Elsevier Ltd
01.06.2021
Elsevier Science Ltd |
| Schlagworte: | |
| ISSN: | 0191-2615, 1879-2367 |
| Online-Zugang: | Volltext |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| Abstract | •Used statistical learning theory to evaluate DNNs in choice analysis.•Operationalized DNN interpretability by using the choice probability functions.•Provided a tight upper bound on the estimation error of DNNs.•Conducted experiments to identify when DNNs outperform classical models.•DNNs can be more predictive and interpretable than BNL and MNL models.
Although researchers increasingly use deep neural networks (DNN) to analyze individual choices, overfitting and interpretability issues remain obstacles in theory and practice. This study presents a statistical learning theoretical framework to examine the tradeoff between estimation and approximation errors, and between the quality of prediction and of interpretation. It provides an upper bound on the estimation error of the prediction quality in DNN, measured by zero-one and log losses, shedding light on why DNN models do not overfit. It proposes a metric for interpretation quality by formulating a function approximation loss that measures the difference between true and estimated choice probability functions. It argues that the binary logit (BNL) and multinomial logit (MNL) models are the specific cases of DNNs, since the latter always has smaller approximation errors. We explore the relative performance of DNN and classical choice models through three simulation scenarios comparing DNN, BNL, and binary mixed logit models (BXL), as well as one experiment comparing DNN to BNL, BXL, MNL, and mixed logit (MXL) in analyzing the choice of trip purposes based on the National Household Travel Survey 2017. The results indicate that DNN can be used for choice analysis beyond the current practice of demand forecasting because it has the inherent utility interpretation and the power of automatically learning utility specification. Our results suggest DNN outperforms BNL, BXL, MNL, and MXL models in both prediction and interpretation when the sample size is large (≥O(104)), the input dimension is high, or the true data generating process is complex, while performing worse when the opposite is true. DNN outperforms BNL and BXL in zero-one, log, and approximation losses for most of the experiments, and the larger sample size leads to greater incremental value of using DNN over classical discrete choice models. Overall, this study introduces the statistical learning theory as a new foundation for high-dimensional data, complex statistical models, and non-asymptotic data regimes in choice analysis, and the experiments show the effective prediction and interpretation of DNN for its applications to policy and behavioral analysis. |
|---|---|
| AbstractList | Although researchers increasingly use deep neural networks (DNN) to analyze individual choices, overfitting and interpretability issues remain obstacles in theory and practice. This study presents a statistical learning theoretical framework to examine the tradeoff between estimation and approximation errors, and between the quality of prediction and of interpretation. It provides an upper bound on the estimation error of the prediction quality in DNN, measured by zero-one and log losses, shedding light on why DNN models do not overfit. It proposes a metric for interpretation quality by formulating a function approximation loss that measures the difference between true and estimated choice probability functions. It argues that the binary logit (BNL) and multinomial logit (MNL) models are the specific cases of DNNs, since the latter always has smaller approximation errors. We explore the relative performance of DNN and classical choice models through three simulation scenarios comparing DNN, BNL, and binary mixed logit models (BXL), as well as one experiment comparing DNN to BNL, BXL, MNL, and mixed logit (MXL) in analyzing the choice of trip purposes based on the National Household Travel Survey 2017. The results indicate that DNN can be used for choice analysis beyond the current practice of demand forecasting because it has the inherent utility interpretation and the power of automatically learning utility specification. Our results suggest DNN outperforms BNL, BXL, MNL, and MXL models in both prediction and interpretation when the sample size is large (≥ O ( 104)), the input dimension is high, or the true data generating process is complex, while performing worse when the opposite is true. DNN outperforms BNL and BXL in zero-one, log, and approximation losses for most of the experiments, and the larger sample size leads to greater incremental value of using DNN over classical discrete choice models. Overall, this study introduces the statistical learning theory as a new foundation for high-dimensional data, complex statistical models, and non-asymptotic data regimes in choice analysis, and the experiments show the effective prediction and interpretation of DNN for its applications to policy and behavioral analysis. •Used statistical learning theory to evaluate DNNs in choice analysis.•Operationalized DNN interpretability by using the choice probability functions.•Provided a tight upper bound on the estimation error of DNNs.•Conducted experiments to identify when DNNs outperform classical models.•DNNs can be more predictive and interpretable than BNL and MNL models. Although researchers increasingly use deep neural networks (DNN) to analyze individual choices, overfitting and interpretability issues remain obstacles in theory and practice. This study presents a statistical learning theoretical framework to examine the tradeoff between estimation and approximation errors, and between the quality of prediction and of interpretation. It provides an upper bound on the estimation error of the prediction quality in DNN, measured by zero-one and log losses, shedding light on why DNN models do not overfit. It proposes a metric for interpretation quality by formulating a function approximation loss that measures the difference between true and estimated choice probability functions. It argues that the binary logit (BNL) and multinomial logit (MNL) models are the specific cases of DNNs, since the latter always has smaller approximation errors. We explore the relative performance of DNN and classical choice models through three simulation scenarios comparing DNN, BNL, and binary mixed logit models (BXL), as well as one experiment comparing DNN to BNL, BXL, MNL, and mixed logit (MXL) in analyzing the choice of trip purposes based on the National Household Travel Survey 2017. The results indicate that DNN can be used for choice analysis beyond the current practice of demand forecasting because it has the inherent utility interpretation and the power of automatically learning utility specification. Our results suggest DNN outperforms BNL, BXL, MNL, and MXL models in both prediction and interpretation when the sample size is large (≥O(104)), the input dimension is high, or the true data generating process is complex, while performing worse when the opposite is true. DNN outperforms BNL and BXL in zero-one, log, and approximation losses for most of the experiments, and the larger sample size leads to greater incremental value of using DNN over classical discrete choice models. Overall, this study introduces the statistical learning theory as a new foundation for high-dimensional data, complex statistical models, and non-asymptotic data regimes in choice analysis, and the experiments show the effective prediction and interpretation of DNN for its applications to policy and behavioral analysis. |
| Author | Wang, Shenhao Zhao, Jinhua Wang, Qingyi Bailey, Nate |
| Author_xml | – sequence: 1 givenname: Shenhao surname: Wang fullname: Wang, Shenhao – sequence: 2 givenname: Qingyi surname: Wang fullname: Wang, Qingyi – sequence: 3 givenname: Nate surname: Bailey fullname: Bailey, Nate – sequence: 4 givenname: Jinhua surname: Zhao fullname: Zhao, Jinhua email: jinhua@mit.edu |
| BookMark | eNp9kD1PwzAURS1UJErhB7BFYk7ws524gQnxLSExALPlOs_UJcTBdkH997gqEwPTXe55evccksngByTkBGgFFJqzVZXComKUQUV5RQH2yBTmsi0Zb-SETCm0ULIG6gNyGOOKUsoFhSl5vkYciwHXQfc50rcP77GwPhRm6Z3BQg-630QXz4vLIiadXEzO5G6POgxueCvSEn3YFCOGOKJJ7guPyL7VfcTj35yR19ubl6v78vHp7uHq8rE0vJ6nkhveCaY1RdQLEDXXYCTTtsNFK7AFqeVcMjOvUVjR2aZhpmMIzULw1kpr-Yyc7u6OwX-uMSa18uuQ_42K1UIwSUVT55bctUzwMQa0yrjtDj-koF2vgKqtQbVS2aDaGlSUq2wwk_CHHIP70GHzL3OxYzAP_3IYVDQOB4OdC9mO6rz7h_4Bp7WMjw |
| CitedBy_id | crossref_primary_10_1016_j_trb_2023_102869 crossref_primary_10_1080_03081060_2025_2520571 crossref_primary_10_1177_03611981241310399 crossref_primary_10_1016_j_trb_2025_103318 crossref_primary_10_1016_j_rineng_2025_105714 crossref_primary_10_1016_j_tre_2025_104154 crossref_primary_10_1049_itr2_12514 crossref_primary_10_1108_JEIM_01_2022_0025 crossref_primary_10_1007_s40864_025_00245_9 crossref_primary_10_1016_j_trc_2021_103410 crossref_primary_10_1080_10298436_2023_2257852 crossref_primary_10_1016_j_eswa_2024_125196 crossref_primary_10_1109_TITS_2022_3216462 crossref_primary_10_1371_journal_pone_0331664 crossref_primary_10_1016_j_trc_2022_103924 crossref_primary_10_3390_rs13183743 crossref_primary_10_1016_j_socscimed_2023_115910 crossref_primary_10_1016_j_trc_2024_104618 crossref_primary_10_1186_s42400_023_00161_0 crossref_primary_10_1016_j_aap_2023_107282 crossref_primary_10_1016_j_trb_2024_103061 crossref_primary_10_1016_j_trb_2022_02_007 crossref_primary_10_1111_joes_70008 crossref_primary_10_1177_03611981231162598 crossref_primary_10_1016_j_trc_2024_104671 crossref_primary_10_3390_app12189156 |
| Cites_doi | 10.1111/ecin.12364 10.1016/j.trc.2017.02.024 10.1007/BF02551274 10.1214/ss/1009213726 10.1016/S1361-9209(97)00009-6 10.1109/72.788640 10.1016/0893-6080(89)90020-8 10.1007/s11633-017-1054-2 10.2307/2296997 10.1016/j.trc.2005.04.002 10.1016/S1366-5545(99)00030-7 10.1016/j.dsp.2017.10.011 10.1016/j.mcm.2006.02.002 10.1016/j.trc.2010.10.004 10.1016/j.trpro.2015.09.037 10.1080/03081060.2015.1079385 10.1016/j.trc.2018.03.001 10.1016/j.trb.2013.09.008 10.1016/j.trc.2020.01.012 10.1016/j.jocm.2020.100236 10.1038/nature14539 10.1016/B978-0-444-52936-7.50016-1 10.1016/0893-6080(91)90009-T 10.1002/(SICI)1099-131X(200004)19:3<177::AID-FOR738>3.0.CO;2-6 10.1016/j.trb.2018.10.020 10.1198/016214506000001437 10.1198/016214505000000907 10.1177/0042098009356125 10.1016/j.eswa.2017.01.057 10.1016/S0191-2615(99)00014-4 10.1016/S0198-9715(98)00036-2 10.1016/j.tbs.2018.09.002 10.1257/jep.31.2.87 10.1016/j.trc.2020.102701 10.1109/TPAMI.2013.50 10.3141/1854-06 |
| ContentType | Journal Article |
| Copyright | 2021 Copyright Elsevier Science Ltd. Jun 2021 |
| Copyright_xml | – notice: 2021 – notice: Copyright Elsevier Science Ltd. Jun 2021 |
| DBID | AAYXX CITATION 7ST 8FD C1K FR3 KR7 SOI |
| DOI | 10.1016/j.trb.2021.03.011 |
| DatabaseName | CrossRef Environment Abstracts Technology Research Database Environmental Sciences and Pollution Management Engineering Research Database Civil Engineering Abstracts Environment Abstracts |
| DatabaseTitle | CrossRef Civil Engineering Abstracts Engineering Research Database Technology Research Database Environment Abstracts Environmental Sciences and Pollution Management |
| DatabaseTitleList | Civil Engineering Abstracts |
| DeliveryMethod | fulltext_linktorsrc |
| Discipline | Economics Engineering |
| EISSN | 1879-2367 |
| EndPage | 81 |
| ExternalDocumentID | 10_1016_j_trb_2021_03_011 S0191261521000564 |
| GroupedDBID | --K --M -~X .DC .~1 0R~ 123 1B1 1RT 1~. 1~5 29Q 4.4 457 4G. 5VS 7-5 71M 8P~ 9JO AAAKF AAAKG AACTN AAEDT AAEDW AAFJI AAIAV AAIKJ AAKOC AALRI AAOAW AAQFI AAQXK AARIN AAXUO ABDEX ABDMP ABFNM ABLJU ABMAC ABMMH ABPPZ ABUCO ABXDB ABYKQ ACDAQ ACGFS ACNCT ACRLP ADBBV ADEZE ADMUD AEBSH AEKER AFKWA AFTJW AGHFR AGUBO AGYEJ AHHHB AHRSL AIEXJ AIKHN AITUG AJBFU AJOXV AKYCK ALMA_UNASSIGNED_HOLDINGS AMFUW AMRAJ AOMHK APLSM ASPBG AVARZ AVWKF AXJTR AZFZN BKOJK BLXMC CS3 EBS EFJIC EFLBG EJD EO8 EO9 EP2 EP3 FDB FEDTE FGOYB FIRID FNPLU FYGXN G-2 G-Q GBLVA HAMUX HMY HVGLF HZ~ H~9 IHE J1W KOM LY1 LY7 M3Y M41 MO0 MS~ N9A O-L O9- OAUVE OHT OZT P-8 P-9 P2P PC. PRBVW Q38 R2- RIG ROL RPZ SDF SDG SDP SDS SES SET SEW SPCBC SSB SSD SSO SSS SSZ T5K WUQ XPP ~G- 9DU AATTM AAXKI AAYWO AAYXX ABWVN ACLOT ACRPL ADNMO AEIPS AFJKZ AGQPQ AIIUN ANKPU APXCP CITATION EFKBS ~HD 7ST 8FD AGCQF C1K FR3 KR7 SOI |
| ID | FETCH-LOGICAL-c358t-3c3d42aa0eeab1453a1c72afdeb94e917a7872c85e4f4df662cd2e16b439f7ff3 |
| ISICitedReferencesCount | 31 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000655548800003&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| ISSN | 0191-2615 |
| IngestDate | Wed Aug 13 11:35:52 EDT 2025 Tue Nov 18 21:26:34 EST 2025 Sat Nov 29 07:25:56 EST 2025 Fri Feb 23 02:45:45 EST 2024 |
| IsPeerReviewed | true |
| IsScholarly | true |
| Keywords | Statistical learning theory Deep neural networks Interpretability Choice modeling |
| Language | English |
| LinkModel | OpenURL |
| MergedId | FETCHMERGED-LOGICAL-c358t-3c3d42aa0eeab1453a1c72afdeb94e917a7872c85e4f4df662cd2e16b439f7ff3 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14 |
| PQID | 2544270465 |
| PQPubID | 2047452 |
| PageCount | 22 |
| ParticipantIDs | proquest_journals_2544270465 crossref_citationtrail_10_1016_j_trb_2021_03_011 crossref_primary_10_1016_j_trb_2021_03_011 elsevier_sciencedirect_doi_10_1016_j_trb_2021_03_011 |
| PublicationCentury | 2000 |
| PublicationDate | June 2021 2021-06-00 20210601 |
| PublicationDateYYYYMMDD | 2021-06-01 |
| PublicationDate_xml | – month: 06 year: 2021 text: June 2021 |
| PublicationDecade | 2020 |
| PublicationPlace | Oxford |
| PublicationPlace_xml | – name: Oxford |
| PublicationTitle | Transportation research. Part B: methodological |
| PublicationYear | 2021 |
| Publisher | Elsevier Ltd Elsevier Science Ltd |
| Publisher_xml | – name: Elsevier Ltd – name: Elsevier Science Ltd |
| References | Vapnik (bib0069) 2013 Neyshabur, Tomioka, Srebro (bib0051) 2015 Wang, Wang, Zhao (bib0074) 2020 Cohen, Ericson, Laibson, White (bib0018) 2016 Vapnik (bib0068) 1999; 10 Breiman (bib0013) 2001; 16 Train (bib0067) 2009 Zhou, Khosla, Lapedriza, Oliva, Torralba (bib0078) 2014 Lipton (bib0046) 2016 Hornik (bib0037) 1991; 4 Karlaftis, Vlahogianni (bib0039) 2011; 19 Cantarella, de Luca (bib0014) 2005; 13 Cervero, Kockelman (bib0016) 1997; 2 Cheng, Chen, De Vos, Lai, Witlox (bib0017) 2019; 14 Von Luxburg, Schölkopf (bib0070) 2011; 10 Bengio, Courville, Vincent (bib0008) 2013; 35 Omrani (bib0052) 2015; 10 Golowich, Rakhlin, Shamir (bib0026) 2017 Bentz, Merunka (bib0009) 2000; 19 Hornik, Stinchcombe, White (bib0038) 1989; 2 Zegras (bib0077) 2010; 47 Bartlett, Mendelson (bib0006) 2002; 3 Train (bib0066) 1980; 47 He, Zhang, Ren, Sun (bib0030) 2015 Gneiting, Raftery (bib0025) 2007; 102 Rolnick, Tegmark (bib0060) 2017 Wainwright (bib0071) 2019; 48 Rao, Sikdar, Rao, Dhingra (bib0058) 1998; 22 Szegedy, Zaremba, Sutskever, Bruna, Erhan, Goodfellow, Fergus (bib0064) 2014 Bertsimas, Delarue, Jaillet, Martin (bib0010) 2019 Doshi-Velez, Kim (bib0022) 2017 Cybenko (bib0019) 1989; 2 Hinton, Vinyals, Dean (bib0036) 2015 Hagenauer, Helbich (bib0028) 2017; 78 Hillel, Bierlaire, Elshafie, Jin (bib0035) 2020 Montavon, Samek, Muller (bib0048) 2018; 73 Ledoux, Talagrand (bib0044) 2013 He, Zhang, Ren, Sun (bib0031) 2016 Wu, Tan, Qin, Ran, Jiang (bib0075) 2018; 90 Mozolin, Thill, Usery (bib0049) 2000; 34 Sontag (bib0062) 1998; 168 Bartlett, Jordan, McAuliffe (bib0005) 2006; 101 Wang, Wang, Zhao (bib0073) 2020; 118 Liao, Poggio (bib0045) 2018 Goodfellow, Bengio, Courville, Bengio (bib0027) 2016; 1 Poggio, Liao, Miranda, Banburski, Boix, Hidary (bib0055) 2018 LeCun, Bengio, Hinton (bib0043) 2015; 521 Mullainathan, Spiess (bib0050) 2017; 31 McFadden (bib0047) 1974 Ribeiro, Singh, Guestrin (bib0059) 2016 Harrell (bib0029) 2015 Baehrens, Schroeter, Harmeling, Kawanabe, Hansen, Müller (bib0003) 2010; 11 Krizhevsky, Sutskever, Hinton (bib0042) 2012 Bousquet, Boucheron, Lugosi (bib0012) 2004 Polson, Sokolov (bib0057) 2017; 79 Glaeser, Kominers, Luca, Naik (bib0024) 2018; 56 Anthony, Bartlett (bib0002) 2009 Hensher, Ton (bib0032) 2000; 36 Poggio, Kawaguchi, Liao, Miranda, Rosasco, Boix, Hidary, Mhaskar (bib0054) 2018 Bishop (bib0011) 2006 Celikoglu (bib0015) 2006; 44 Kingma, Ba (bib0040) 2014 Kotsiantis, Zaharakis, Pintelas (bib0041) 2007; 160 Poggio, Mhaskar, Rosasco, Miranda, Liao (bib0056) 2017; 14 Allahviranloo, Recker (bib0001) 2013; 58 Hillel, Bierlaire, Elshafie, Jin (bib0034) 2019 Paredes, Hemberg, O’Reilly, Zegras (bib0053) 2017 Fernández-Delgado, Cernadas, Barro, Amorim (bib0023) 2014; 15 Dong, Shao, Clarke, Nambisan (bib0021) 2018; 118 Tang, Xiong, Zhang (bib0065) 2015; 38 Soudry, Hoffer, Nacson, Gunasekar, Srebro (bib0063) 2018; 19 Hillel (bib0033) 2020 Xie, Lu, Parkany (bib0076) 2003 De Dios Ortuzar, Willumsen (bib0020) 2011 Ross, Doshi-Velez (bib0061) 2018 Wang, Mo, Zhao (bib0072) 2020; 112 Ben-Akiva, Lerman (bib0007) 1985; 9 Bartlett, Harvey, Liaw, Mehrabian (bib0004) 2017 Liao (10.1016/j.trb.2021.03.011_bib0045) 2018 Zegras (10.1016/j.trb.2021.03.011_bib0077) 2010; 47 Krizhevsky (10.1016/j.trb.2021.03.011_bib0042) 2012 McFadden (10.1016/j.trb.2021.03.011_bib0047) 1974 Hagenauer (10.1016/j.trb.2021.03.011_bib0028) 2017; 78 Doshi-Velez (10.1016/j.trb.2021.03.011_bib0022) 2017 Rolnick (10.1016/j.trb.2021.03.011_bib0060) 2017 Cantarella (10.1016/j.trb.2021.03.011_bib0014) 2005; 13 Poggio (10.1016/j.trb.2021.03.011_bib0055) 2018 Sontag (10.1016/j.trb.2021.03.011_bib0062) 1998; 168 Tang (10.1016/j.trb.2021.03.011_bib0065) 2015; 38 Vapnik (10.1016/j.trb.2021.03.011_bib0068) 1999; 10 Train (10.1016/j.trb.2021.03.011_bib0067) 2009 Ribeiro (10.1016/j.trb.2021.03.011_bib0059) 2016 Bartlett (10.1016/j.trb.2021.03.011_bib0004) 2017 Wainwright (10.1016/j.trb.2021.03.011_bib0071) 2019; 48 Wang (10.1016/j.trb.2021.03.011_bib0074) 2020 LeCun (10.1016/j.trb.2021.03.011_bib0043) 2015; 521 Hornik (10.1016/j.trb.2021.03.011_bib0037) 1991; 4 Fernández-Delgado (10.1016/j.trb.2021.03.011_bib0023) 2014; 15 Montavon (10.1016/j.trb.2021.03.011_bib0048) 2018; 73 Rao (10.1016/j.trb.2021.03.011_bib0058) 1998; 22 Vapnik (10.1016/j.trb.2021.03.011_bib0069) 2013 Train (10.1016/j.trb.2021.03.011_bib0066) 1980; 47 Hensher (10.1016/j.trb.2021.03.011_bib0032) 2000; 36 Omrani (10.1016/j.trb.2021.03.011_bib0052) 2015; 10 Harrell (10.1016/j.trb.2021.03.011_bib0029) 2015 Ross (10.1016/j.trb.2021.03.011_bib0061) 2018 Kingma (10.1016/j.trb.2021.03.011_bib0040) 2014 Hornik (10.1016/j.trb.2021.03.011_bib0038) 1989; 2 Wang (10.1016/j.trb.2021.03.011_bib0072) 2020; 112 Wu (10.1016/j.trb.2021.03.011_bib0075) 2018; 90 He (10.1016/j.trb.2021.03.011_bib0030) 2015 Mozolin (10.1016/j.trb.2021.03.011_bib0049) 2000; 34 Ledoux (10.1016/j.trb.2021.03.011_bib0044) 2013 Hillel (10.1016/j.trb.2021.03.011_bib0034) 2019 Glaeser (10.1016/j.trb.2021.03.011_bib0024) 2018; 56 Bousquet (10.1016/j.trb.2021.03.011_bib0012) 2004 Hinton (10.1016/j.trb.2021.03.011_bib0036) 2015 Wang (10.1016/j.trb.2021.03.011_bib0073) 2020; 118 De Dios Ortuzar (10.1016/j.trb.2021.03.011_bib0020) 2011 Gneiting (10.1016/j.trb.2021.03.011_bib0025) 2007; 102 Poggio (10.1016/j.trb.2021.03.011_bib0056) 2017; 14 Cheng (10.1016/j.trb.2021.03.011_bib0017) 2019; 14 Lipton (10.1016/j.trb.2021.03.011_bib0046) 2016 Xie (10.1016/j.trb.2021.03.011_bib0076) 2003 Dong (10.1016/j.trb.2021.03.011_bib0021) 2018; 118 Cervero (10.1016/j.trb.2021.03.011_bib0016) 1997; 2 Cybenko (10.1016/j.trb.2021.03.011_bib0019) 1989; 2 Allahviranloo (10.1016/j.trb.2021.03.011_bib0001) 2013; 58 Bentz (10.1016/j.trb.2021.03.011_bib0009) 2000; 19 Kotsiantis (10.1016/j.trb.2021.03.011_bib0041) 2007; 160 Ben-Akiva (10.1016/j.trb.2021.03.011_bib0007) 1985; 9 He (10.1016/j.trb.2021.03.011_bib0031) 2016 Bartlett (10.1016/j.trb.2021.03.011_bib0005) 2006; 101 Hillel (10.1016/j.trb.2021.03.011_bib0033) 2020 Szegedy (10.1016/j.trb.2021.03.011_bib0064) 2014 Bishop (10.1016/j.trb.2021.03.011_bib0011) 2006 Soudry (10.1016/j.trb.2021.03.011_bib0063) 2018; 19 Von Luxburg (10.1016/j.trb.2021.03.011_bib0070) 2011; 10 Zhou (10.1016/j.trb.2021.03.011_bib0078) 2014 Poggio (10.1016/j.trb.2021.03.011_bib0054) 2018 Paredes (10.1016/j.trb.2021.03.011_bib0053) 2017 Bertsimas (10.1016/j.trb.2021.03.011_bib0010) 2019 Hillel (10.1016/j.trb.2021.03.011_bib0035) 2020 Baehrens (10.1016/j.trb.2021.03.011_bib0003) 2010; 11 Cohen (10.1016/j.trb.2021.03.011_bib0018) 2016 Breiman (10.1016/j.trb.2021.03.011_bib0013) 2001; 16 Golowich (10.1016/j.trb.2021.03.011_bib0026) 2017 Polson (10.1016/j.trb.2021.03.011_bib0057) 2017; 79 Goodfellow (10.1016/j.trb.2021.03.011_bib0027) 2016; 1 Neyshabur (10.1016/j.trb.2021.03.011_bib0051) 2015 Celikoglu (10.1016/j.trb.2021.03.011_bib0015) 2006; 44 Mullainathan (10.1016/j.trb.2021.03.011_bib0050) 2017; 31 Bartlett (10.1016/j.trb.2021.03.011_bib0006) 2002; 3 Anthony (10.1016/j.trb.2021.03.011_bib0002) 2009 Bengio (10.1016/j.trb.2021.03.011_bib0008) 2013; 35 Karlaftis (10.1016/j.trb.2021.03.011_bib0039) 2011; 19 |
| References_xml | – volume: 47 start-page: 1793 year: 2010 end-page: 1817 ident: bib0077 article-title: The built environment and motor vehicle ownership and use: evidence from santiago de chile publication-title: Urban Studies – year: 2014 ident: bib0040 article-title: Adam: a method for stochastic optimization publication-title: arXiv preprint arXiv:1412.6980 – volume: 118 start-page: 102701 year: 2020 ident: bib0073 article-title: Deep neural networks for choice analysis: extracting complete economic information for interpretation publication-title: Transportation Research Part C: Emerging Technologies – start-page: 100236 year: 2020 ident: bib0074 article-title: Multitask learning deep neural networks to combine revealed and stated preference data publication-title: Journal of Choice Modelling – year: 2019 ident: bib0010 article-title: The price of interpretability publication-title: Arxiv preprint – year: 2014 ident: bib0064 article-title: Intriguing properties of neural networks publication-title: arXiv preprint arXiv:1312.6199 – volume: 79 start-page: 1 year: 2017 end-page: 17 ident: bib0057 article-title: Deep learning for short-term traffic flow prediction publication-title: Transportation Research Part C: Emerging Technologies – volume: 2 start-page: 199 year: 1997 end-page: 219 ident: bib0016 article-title: Travel demand and the 3ds: density, diversity, and design publication-title: Transportation Research Part D: Transport and Environment – year: 2018 ident: bib0045 article-title: When Is Handcrafting Not a Curse? publication-title: Technical Report – year: 1974 ident: bib0047 article-title: Conditional Logit Analysis of Qualitative Choice Behavior – start-page: 1376 year: 2015 end-page: 1401 ident: bib0051 article-title: Norm-based capacity control in neural networks publication-title: Conference on Learning Theory – start-page: 50 year: 2003 end-page: 61 ident: bib0076 article-title: Work travel mode choice modeling with data mining: decision trees and neural networks publication-title: Transportation Research Record: Journal of the Transportation Research Board – volume: 9 year: 1985 ident: bib0007 article-title: Discrete choice analysis: Theory and application to travel demand – volume: 14 start-page: 1 year: 2019 end-page: 10 ident: bib0017 article-title: Applying a random forest method approach to model travel mode choice behavior publication-title: Travel behaviour and society – year: 2006 ident: bib0011 article-title: Pattern recognition and machine learning – year: 2011 ident: bib0020 article-title: Modelling transport – volume: 521 start-page: 436 year: 2015 end-page: 444 ident: bib0043 article-title: Deep learning publication-title: Nature – volume: 16 start-page: 199 year: 2001 end-page: 231 ident: bib0013 article-title: Statistical modeling: the two cultures (with comments and a rejoinder by the author) publication-title: Statistical science – volume: 14 start-page: 503 year: 2017 end-page: 519 ident: bib0056 article-title: Why and when can deep-but not shallow-networks avoid the curse of dimensionality: a review publication-title: Int. J. Autom. Comput. – year: 2020 ident: bib0033 article-title: New Perspectives on the Performance of Machine Learning Classifiers for Mode Choice Prediction – year: 2018 ident: bib0061 article-title: Improving the adversarial robustness and interpretability of deep neural networks by regularizing their input gradients publication-title: Thirty-second AAAI conference on artificial intelligence – year: 2013 ident: bib0069 article-title: The nature of statistical learning theory – volume: 2 start-page: 359 year: 1989 end-page: 366 ident: bib0038 article-title: Multilayer feedforward networks are universal approximators publication-title: Neural networks – start-page: 1026 year: 2015 end-page: 1034 ident: bib0030 article-title: Delving deep into rectifiers: Surpassing human-level performance on imagenet classification publication-title: Proceedings of the IEEE international conference on computer vision – volume: 73 start-page: 1 year: 2018 end-page: 15 ident: bib0048 article-title: Methods for interpreting and understanding deep neural networks publication-title: Digit. Signal Process. – volume: 112 start-page: 234 year: 2020 end-page: 251 ident: bib0072 article-title: Deep neural networks for choice analysis: architecture design with alternative-specific utility functions publication-title: Transportation Research Part C: Emerging Technologies – volume: 19 start-page: 2822 year: 2018 end-page: 2878 ident: bib0063 article-title: The implicit bias of gradient descent on separable data publication-title: The Journal of Machine Learning Research – start-page: 1097 year: 2012 end-page: 1105 ident: bib0042 article-title: Imagenet classification with deep convolutional neural networks publication-title: Advances in neural information processing systems – volume: 22 start-page: 485 year: 1998 end-page: 496 ident: bib0058 article-title: Another insight into artificial neural networks through behavioural analysis of access mode choice publication-title: Comput. Environ. Urban Syst. – volume: 38 start-page: 833 year: 2015 end-page: 850 ident: bib0065 article-title: Decision tree method for modeling travel mode switching in a dynamic behavioral process publication-title: Transportation Planning and Technology – volume: 58 start-page: 16 year: 2013 end-page: 43 ident: bib0001 article-title: Daily activity pattern recognition by using support vector machines with multiple classes publication-title: Transportation Research Part B: Methodological – volume: 13 start-page: 121 year: 2005 end-page: 155 ident: bib0014 article-title: Multilayer feedforward networks for transportation mode choice analysis: an analysis and a comparison with random utility models publication-title: Transportation Research Part C: Emerging Technologies – volume: 78 start-page: 273 year: 2017 end-page: 282 ident: bib0028 article-title: A comparative study of machine learning classifiers for modeling travel mode choice publication-title: Expert Syst. Appl. – start-page: 780 year: 2017 end-page: 785 ident: bib0053 article-title: Machine learning or discrete choice models for car ownership demand estimation and prediction? publication-title: Models and Technologies for Intelligent Transportation Systems (MT-ITS), 2017 5th IEEE International Conference on – year: 2019 ident: bib0034 article-title: Weak teachers: Assisted specification of discrete choice models using ensemble learning publication-title: 8th Symposium of the European Association for Research in Transportation, Budapest – volume: 10 start-page: 651 year: 2011 end-page: 706 ident: bib0070 article-title: Statistical Learning Theory: Models, Concepts, and Results publication-title: Handbook of the History of Logic – volume: 118 start-page: 407 year: 2018 end-page: 428 ident: bib0021 article-title: An innovative approach for traffic crash estimation and prediction on accommodating unobserved heterogeneities publication-title: Transportation research part B: methodological – volume: 56 start-page: 114 year: 2018 end-page: 137 ident: bib0024 article-title: Big data and big cities: the promises and limitations of improved measures of urban life publication-title: Econ. Inq. – year: 2015 ident: bib0036 article-title: Distilling the knowledge in a neural network publication-title: arXiv preprint arXiv:1503.02531 – volume: 19 start-page: 177 year: 2000 end-page: 200 ident: bib0009 article-title: Neural networks and the multinomial logit for brand choice modelling: a hybrid approach publication-title: J. Forecast. – start-page: 169 year: 2004 end-page: 207 ident: bib0012 article-title: Introduction to Statistical Learning Theory publication-title: Advanced lectures on machine learning – year: 2017 ident: bib0022 article-title: Towards a Rigorous Science of Interpretable Machine Learning – year: 2017 ident: bib0026 article-title: Size-independent sample complexity of neural networks publication-title: arXiv preprint arXiv:1712.06541 – year: 2016 ident: bib0046 article-title: The mythos of model interpretability publication-title: arXiv preprint arXiv:1606.03490 – start-page: 770 year: 2016 end-page: 778 ident: bib0031 article-title: Deep residual learning for image recognition publication-title: Proceedings of the IEEE conference on computer vision and pattern recognition – year: 2018 ident: bib0054 article-title: Theory of deep learning iii: the non-overfitting puzzle publication-title: Technical Report – volume: 36 start-page: 155 year: 2000 end-page: 172 ident: bib0032 article-title: A comparison of the predictive potential of artificial neural networks and nested logit models for commuter mode choice publication-title: Transportation Research Part E: Logistics and Transportation Review – volume: 2 start-page: 303 year: 1989 end-page: 314 ident: bib0019 article-title: Approximation by superpositions of a sigmoidal function publication-title: Mathematics of control, signals and systems – year: 2016 ident: bib0018 article-title: Measuring time preferences publication-title: Technical Report – volume: 31 start-page: 87 year: 2017 end-page: 106 ident: bib0050 article-title: Machine learning: an applied econometric approach publication-title: Journal of Economic Perspectives – volume: 10 start-page: 840 year: 2015 end-page: 849 ident: bib0052 article-title: Predicting travel mode of individuals by machine learning publication-title: Transp. Res. Procedia – volume: 19 start-page: 387 year: 2011 end-page: 399 ident: bib0039 article-title: Statistical methods versus neural networks in transportation research: differences, similarities and some insights publication-title: Transportation Research Part C: Emerging Technologies – volume: 44 start-page: 640 year: 2006 end-page: 658 ident: bib0015 article-title: Application of radial basis function and generalized regression neural networks in non-linear utility function specification for travel mode choice modelling publication-title: Math. Comput. Model. – volume: 10 start-page: 988 year: 1999 end-page: 999 ident: bib0068 article-title: An overview of statistical learning theory publication-title: IEEE Trans. Neural Networks – volume: 102 start-page: 359 year: 2007 end-page: 378 ident: bib0025 article-title: Strictly proper scoring rules, prediction, and estimation publication-title: J. Am. Stat. Assoc. – year: 2009 ident: bib0067 article-title: Discrete choice methods with simulation – volume: 101 start-page: 138 year: 2006 end-page: 156 ident: bib0005 article-title: Convexity, classification, and risk bounds publication-title: J. Am. Stat. Assoc. – start-page: 1135 year: 2016 end-page: 1144 ident: bib0059 article-title: Why should i trust you?: Explaining the predictions of any classifier publication-title: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining – year: 2013 ident: bib0044 article-title: Probability in banach spaces: Isoperimetry and processes – year: 2017 ident: bib0060 article-title: The power of deeper networks for expressing natural functions publication-title: arXiv preprint arXiv:1705.05502 – volume: 168 start-page: 69 year: 1998 end-page: 96 ident: bib0062 article-title: Vc dimension of neural networks publication-title: NATO ASI Series F Computer and Systems Sciences – volume: 1 year: 2016 ident: bib0027 article-title: Deep learning – year: 2017 ident: bib0004 article-title: Nearly-tight vc-dimension and pseudodimension bounds for piecewise linear neural networks publication-title: arXiv preprint arXiv:1703.02930 – volume: 4 start-page: 251 year: 1991 end-page: 257 ident: bib0037 article-title: Approximation capabilities of multilayer feedforward networks publication-title: Neural networks – start-page: 100221 year: 2020 ident: bib0035 article-title: A systematic review of machine learning classification methodologies for modelling passenger mode choice publication-title: Journal of Choice Modelling – volume: 48 year: 2019 ident: bib0071 article-title: High-dimensional statistics: A non-asymptotic viewpoint – year: 2009 ident: bib0002 article-title: Neural network learning: Theoretical foundations – volume: 11 start-page: 1803 year: 2010 end-page: 1831 ident: bib0003 article-title: How to explain individual classification decisions publication-title: Journal of Machine Learning Research – volume: 34 start-page: 53 year: 2000 end-page: 73 ident: bib0049 article-title: Trip distribution forecasting with multilayer perceptron neural networks: a critical evaluation publication-title: Transportation Research Part B: Methodological – volume: 15 start-page: 3133 year: 2014 end-page: 3181 ident: bib0023 article-title: Do we need hundreds of classifiers to solve real world classification problems publication-title: Journal of Machine Learning Research – volume: 47 start-page: 357 year: 1980 end-page: 370 ident: bib0066 article-title: A structured logit model of auto ownership and mode choice publication-title: Rev Econ Stud – year: 2014 ident: bib0078 article-title: Object detectors emerge in deep scene cnns publication-title: arXiv preprint arXiv:1412.6856 – year: 2015 ident: bib0029 article-title: Regression modeling strategies: With applications to linear models, logistic and ordinal regression, and survival analysis – year: 2018 ident: bib0055 article-title: Theory iiib: generalization in deep networks publication-title: arXiv preprint arXiv:1806.11379 – volume: 35 start-page: 1798 year: 2013 end-page: 1828 ident: bib0008 article-title: Representation learning: a review and new perspectives publication-title: IEEE Trans. Pattern Anal. Mach. Intell. – volume: 160 start-page: 3 year: 2007 end-page: 24 ident: bib0041 article-title: Supervised machine learning: a review of classification techniques publication-title: Emerging artificial intelligence applications in computer engineering – volume: 90 start-page: 166 year: 2018 end-page: 180 ident: bib0075 article-title: A hybrid deep learning based traffic flow prediction method and its understanding publication-title: Transportation Research Part C: Emerging Technologies – volume: 3 start-page: 463 year: 2002 end-page: 482 ident: bib0006 article-title: Rademacher and gaussian complexities: risk bounds and structural results publication-title: Journal of Machine Learning Research – volume: 56 start-page: 114 issue: 1 year: 2018 ident: 10.1016/j.trb.2021.03.011_bib0024 article-title: Big data and big cities: the promises and limitations of improved measures of urban life publication-title: Econ. Inq. doi: 10.1111/ecin.12364 – start-page: 1026 year: 2015 ident: 10.1016/j.trb.2021.03.011_bib0030 article-title: Delving deep into rectifiers: Surpassing human-level performance on imagenet classification – year: 2019 ident: 10.1016/j.trb.2021.03.011_bib0010 article-title: The price of interpretability publication-title: Arxiv preprint – volume: 79 start-page: 1 year: 2017 ident: 10.1016/j.trb.2021.03.011_bib0057 article-title: Deep learning for short-term traffic flow prediction publication-title: Transportation Research Part C: Emerging Technologies doi: 10.1016/j.trc.2017.02.024 – volume: 19 start-page: 2822 issue: 1 year: 2018 ident: 10.1016/j.trb.2021.03.011_bib0063 article-title: The implicit bias of gradient descent on separable data publication-title: The Journal of Machine Learning Research – volume: 2 start-page: 303 issue: 4 year: 1989 ident: 10.1016/j.trb.2021.03.011_bib0019 article-title: Approximation by superpositions of a sigmoidal function publication-title: Mathematics of control, signals and systems doi: 10.1007/BF02551274 – volume: 16 start-page: 199 issue: 3 year: 2001 ident: 10.1016/j.trb.2021.03.011_bib0013 article-title: Statistical modeling: the two cultures (with comments and a rejoinder by the author) publication-title: Statistical science doi: 10.1214/ss/1009213726 – volume: 2 start-page: 199 issue: 3 year: 1997 ident: 10.1016/j.trb.2021.03.011_bib0016 article-title: Travel demand and the 3ds: density, diversity, and design publication-title: Transportation Research Part D: Transport and Environment doi: 10.1016/S1361-9209(97)00009-6 – start-page: 100221 year: 2020 ident: 10.1016/j.trb.2021.03.011_bib0035 article-title: A systematic review of machine learning classification methodologies for modelling passenger mode choice publication-title: Journal of Choice Modelling – volume: 10 start-page: 988 issue: 5 year: 1999 ident: 10.1016/j.trb.2021.03.011_bib0068 article-title: An overview of statistical learning theory publication-title: IEEE Trans. Neural Networks doi: 10.1109/72.788640 – year: 2016 ident: 10.1016/j.trb.2021.03.011_bib0018 article-title: Measuring time preferences – start-page: 770 year: 2016 ident: 10.1016/j.trb.2021.03.011_bib0031 article-title: Deep residual learning for image recognition – volume: 2 start-page: 359 issue: 5 year: 1989 ident: 10.1016/j.trb.2021.03.011_bib0038 article-title: Multilayer feedforward networks are universal approximators publication-title: Neural networks doi: 10.1016/0893-6080(89)90020-8 – volume: 14 start-page: 503 issue: 5 year: 2017 ident: 10.1016/j.trb.2021.03.011_bib0056 article-title: Why and when can deep-but not shallow-networks avoid the curse of dimensionality: a review publication-title: Int. J. Autom. Comput. doi: 10.1007/s11633-017-1054-2 – volume: 47 start-page: 357 issue: 2 year: 1980 ident: 10.1016/j.trb.2021.03.011_bib0066 article-title: A structured logit model of auto ownership and mode choice publication-title: Rev Econ Stud doi: 10.2307/2296997 – volume: 11 start-page: 1803 issue: Jun year: 2010 ident: 10.1016/j.trb.2021.03.011_bib0003 article-title: How to explain individual classification decisions publication-title: Journal of Machine Learning Research – volume: 13 start-page: 121 issue: 2 year: 2005 ident: 10.1016/j.trb.2021.03.011_bib0014 article-title: Multilayer feedforward networks for transportation mode choice analysis: an analysis and a comparison with random utility models publication-title: Transportation Research Part C: Emerging Technologies doi: 10.1016/j.trc.2005.04.002 – volume: 36 start-page: 155 issue: 3 year: 2000 ident: 10.1016/j.trb.2021.03.011_bib0032 article-title: A comparison of the predictive potential of artificial neural networks and nested logit models for commuter mode choice publication-title: Transportation Research Part E: Logistics and Transportation Review doi: 10.1016/S1366-5545(99)00030-7 – year: 2018 ident: 10.1016/j.trb.2021.03.011_bib0061 article-title: Improving the adversarial robustness and interpretability of deep neural networks by regularizing their input gradients – year: 2017 ident: 10.1016/j.trb.2021.03.011_bib0022 – year: 2018 ident: 10.1016/j.trb.2021.03.011_bib0054 article-title: Theory of deep learning iii: the non-overfitting puzzle – volume: 48 year: 2019 ident: 10.1016/j.trb.2021.03.011_bib0071 – year: 2014 ident: 10.1016/j.trb.2021.03.011_bib0040 article-title: Adam: a method for stochastic optimization publication-title: arXiv preprint arXiv:1412.6980 – volume: 73 start-page: 1 year: 2018 ident: 10.1016/j.trb.2021.03.011_bib0048 article-title: Methods for interpreting and understanding deep neural networks publication-title: Digit. Signal Process. doi: 10.1016/j.dsp.2017.10.011 – volume: 44 start-page: 640 issue: 7 year: 2006 ident: 10.1016/j.trb.2021.03.011_bib0015 article-title: Application of radial basis function and generalized regression neural networks in non-linear utility function specification for travel mode choice modelling publication-title: Math. Comput. Model. doi: 10.1016/j.mcm.2006.02.002 – year: 2020 ident: 10.1016/j.trb.2021.03.011_bib0033 – volume: 19 start-page: 387 issue: 3 year: 2011 ident: 10.1016/j.trb.2021.03.011_bib0039 article-title: Statistical methods versus neural networks in transportation research: differences, similarities and some insights publication-title: Transportation Research Part C: Emerging Technologies doi: 10.1016/j.trc.2010.10.004 – year: 2009 ident: 10.1016/j.trb.2021.03.011_bib0067 – volume: 15 start-page: 3133 issue: 1 year: 2014 ident: 10.1016/j.trb.2021.03.011_bib0023 article-title: Do we need hundreds of classifiers to solve real world classification problems publication-title: Journal of Machine Learning Research – volume: 10 start-page: 840 year: 2015 ident: 10.1016/j.trb.2021.03.011_bib0052 article-title: Predicting travel mode of individuals by machine learning publication-title: Transp. Res. Procedia doi: 10.1016/j.trpro.2015.09.037 – volume: 38 start-page: 833 issue: 8 year: 2015 ident: 10.1016/j.trb.2021.03.011_bib0065 article-title: Decision tree method for modeling travel mode switching in a dynamic behavioral process publication-title: Transportation Planning and Technology doi: 10.1080/03081060.2015.1079385 – volume: 90 start-page: 166 year: 2018 ident: 10.1016/j.trb.2021.03.011_bib0075 article-title: A hybrid deep learning based traffic flow prediction method and its understanding publication-title: Transportation Research Part C: Emerging Technologies doi: 10.1016/j.trc.2018.03.001 – year: 2006 ident: 10.1016/j.trb.2021.03.011_bib0011 – start-page: 1135 year: 2016 ident: 10.1016/j.trb.2021.03.011_bib0059 article-title: Why should i trust you?: Explaining the predictions of any classifier – year: 2017 ident: 10.1016/j.trb.2021.03.011_bib0026 article-title: Size-independent sample complexity of neural networks publication-title: arXiv preprint arXiv:1712.06541 – start-page: 1097 year: 2012 ident: 10.1016/j.trb.2021.03.011_bib0042 article-title: Imagenet classification with deep convolutional neural networks – volume: 58 start-page: 16 year: 2013 ident: 10.1016/j.trb.2021.03.011_bib0001 article-title: Daily activity pattern recognition by using support vector machines with multiple classes publication-title: Transportation Research Part B: Methodological doi: 10.1016/j.trb.2013.09.008 – volume: 112 start-page: 234 year: 2020 ident: 10.1016/j.trb.2021.03.011_bib0072 article-title: Deep neural networks for choice analysis: architecture design with alternative-specific utility functions publication-title: Transportation Research Part C: Emerging Technologies doi: 10.1016/j.trc.2020.01.012 – start-page: 169 year: 2004 ident: 10.1016/j.trb.2021.03.011_bib0012 article-title: Introduction to Statistical Learning Theory – start-page: 100236 year: 2020 ident: 10.1016/j.trb.2021.03.011_bib0074 article-title: Multitask learning deep neural networks to combine revealed and stated preference data publication-title: Journal of Choice Modelling doi: 10.1016/j.jocm.2020.100236 – year: 2015 ident: 10.1016/j.trb.2021.03.011_bib0029 – volume: 521 start-page: 436 issue: 7553 year: 2015 ident: 10.1016/j.trb.2021.03.011_bib0043 article-title: Deep learning publication-title: Nature doi: 10.1038/nature14539 – volume: 10 start-page: 651 year: 2011 ident: 10.1016/j.trb.2021.03.011_bib0070 article-title: Statistical Learning Theory: Models, Concepts, and Results doi: 10.1016/B978-0-444-52936-7.50016-1 – volume: 4 start-page: 251 issue: 2 year: 1991 ident: 10.1016/j.trb.2021.03.011_bib0037 article-title: Approximation capabilities of multilayer feedforward networks publication-title: Neural networks doi: 10.1016/0893-6080(91)90009-T – year: 2009 ident: 10.1016/j.trb.2021.03.011_bib0002 – year: 2016 ident: 10.1016/j.trb.2021.03.011_bib0046 article-title: The mythos of model interpretability publication-title: arXiv preprint arXiv:1606.03490 – volume: 19 start-page: 177 issue: 3 year: 2000 ident: 10.1016/j.trb.2021.03.011_bib0009 article-title: Neural networks and the multinomial logit for brand choice modelling: a hybrid approach publication-title: J. Forecast. doi: 10.1002/(SICI)1099-131X(200004)19:3<177::AID-FOR738>3.0.CO;2-6 – volume: 118 start-page: 407 year: 2018 ident: 10.1016/j.trb.2021.03.011_bib0021 article-title: An innovative approach for traffic crash estimation and prediction on accommodating unobserved heterogeneities publication-title: Transportation research part B: methodological doi: 10.1016/j.trb.2018.10.020 – volume: 102 start-page: 359 issue: 477 year: 2007 ident: 10.1016/j.trb.2021.03.011_bib0025 article-title: Strictly proper scoring rules, prediction, and estimation publication-title: J. Am. Stat. Assoc. doi: 10.1198/016214506000001437 – year: 1974 ident: 10.1016/j.trb.2021.03.011_bib0047 – volume: 101 start-page: 138 issue: 473 year: 2006 ident: 10.1016/j.trb.2021.03.011_bib0005 article-title: Convexity, classification, and risk bounds publication-title: J. Am. Stat. Assoc. doi: 10.1198/016214505000000907 – year: 2013 ident: 10.1016/j.trb.2021.03.011_bib0044 – year: 2014 ident: 10.1016/j.trb.2021.03.011_bib0064 article-title: Intriguing properties of neural networks publication-title: arXiv preprint arXiv:1312.6199 – year: 2014 ident: 10.1016/j.trb.2021.03.011_bib0078 article-title: Object detectors emerge in deep scene cnns publication-title: arXiv preprint arXiv:1412.6856 – year: 2011 ident: 10.1016/j.trb.2021.03.011_bib0020 – start-page: 1376 year: 2015 ident: 10.1016/j.trb.2021.03.011_bib0051 article-title: Norm-based capacity control in neural networks – volume: 47 start-page: 1793 issue: 8 year: 2010 ident: 10.1016/j.trb.2021.03.011_bib0077 article-title: The built environment and motor vehicle ownership and use: evidence from santiago de chile publication-title: Urban Studies doi: 10.1177/0042098009356125 – year: 2013 ident: 10.1016/j.trb.2021.03.011_bib0069 – volume: 78 start-page: 273 year: 2017 ident: 10.1016/j.trb.2021.03.011_bib0028 article-title: A comparative study of machine learning classifiers for modeling travel mode choice publication-title: Expert Syst. Appl. doi: 10.1016/j.eswa.2017.01.057 – year: 2019 ident: 10.1016/j.trb.2021.03.011_bib0034 article-title: Weak teachers: Assisted specification of discrete choice models using ensemble learning – volume: 34 start-page: 53 issue: 1 year: 2000 ident: 10.1016/j.trb.2021.03.011_bib0049 article-title: Trip distribution forecasting with multilayer perceptron neural networks: a critical evaluation publication-title: Transportation Research Part B: Methodological doi: 10.1016/S0191-2615(99)00014-4 – year: 2018 ident: 10.1016/j.trb.2021.03.011_bib0055 article-title: Theory iiib: generalization in deep networks publication-title: arXiv preprint arXiv:1806.11379 – volume: 22 start-page: 485 issue: 5 year: 1998 ident: 10.1016/j.trb.2021.03.011_bib0058 article-title: Another insight into artificial neural networks through behavioural analysis of access mode choice publication-title: Comput. Environ. Urban Syst. doi: 10.1016/S0198-9715(98)00036-2 – volume: 1 year: 2016 ident: 10.1016/j.trb.2021.03.011_bib0027 – volume: 168 start-page: 69 year: 1998 ident: 10.1016/j.trb.2021.03.011_bib0062 article-title: Vc dimension of neural networks publication-title: NATO ASI Series F Computer and Systems Sciences – start-page: 780 year: 2017 ident: 10.1016/j.trb.2021.03.011_bib0053 article-title: Machine learning or discrete choice models for car ownership demand estimation and prediction? – year: 2018 ident: 10.1016/j.trb.2021.03.011_bib0045 article-title: When Is Handcrafting Not a Curse? – volume: 14 start-page: 1 year: 2019 ident: 10.1016/j.trb.2021.03.011_bib0017 article-title: Applying a random forest method approach to model travel mode choice behavior publication-title: Travel behaviour and society doi: 10.1016/j.tbs.2018.09.002 – volume: 9 year: 1985 ident: 10.1016/j.trb.2021.03.011_bib0007 – volume: 160 start-page: 3 year: 2007 ident: 10.1016/j.trb.2021.03.011_bib0041 article-title: Supervised machine learning: a review of classification techniques publication-title: Emerging artificial intelligence applications in computer engineering – volume: 3 start-page: 463 issue: Nov year: 2002 ident: 10.1016/j.trb.2021.03.011_bib0006 article-title: Rademacher and gaussian complexities: risk bounds and structural results publication-title: Journal of Machine Learning Research – year: 2017 ident: 10.1016/j.trb.2021.03.011_bib0060 article-title: The power of deeper networks for expressing natural functions publication-title: arXiv preprint arXiv:1705.05502 – volume: 31 start-page: 87 issue: 2 year: 2017 ident: 10.1016/j.trb.2021.03.011_bib0050 article-title: Machine learning: an applied econometric approach publication-title: Journal of Economic Perspectives doi: 10.1257/jep.31.2.87 – volume: 118 start-page: 102701 year: 2020 ident: 10.1016/j.trb.2021.03.011_bib0073 article-title: Deep neural networks for choice analysis: extracting complete economic information for interpretation publication-title: Transportation Research Part C: Emerging Technologies doi: 10.1016/j.trc.2020.102701 – volume: 35 start-page: 1798 issue: 8 year: 2013 ident: 10.1016/j.trb.2021.03.011_bib0008 article-title: Representation learning: a review and new perspectives publication-title: IEEE Trans. Pattern Anal. Mach. Intell. doi: 10.1109/TPAMI.2013.50 – year: 2017 ident: 10.1016/j.trb.2021.03.011_bib0004 article-title: Nearly-tight vc-dimension and pseudodimension bounds for piecewise linear neural networks publication-title: arXiv preprint arXiv:1703.02930 – year: 2015 ident: 10.1016/j.trb.2021.03.011_bib0036 article-title: Distilling the knowledge in a neural network publication-title: arXiv preprint arXiv:1503.02531 – start-page: 50 issue: 1854 year: 2003 ident: 10.1016/j.trb.2021.03.011_bib0076 article-title: Work travel mode choice modeling with data mining: decision trees and neural networks publication-title: Transportation Research Record: Journal of the Transportation Research Board doi: 10.3141/1854-06 |
| SSID | ssj0003401 |
| Score | 2.5240865 |
| Snippet | •Used statistical learning theory to evaluate DNNs in choice analysis.•Operationalized DNN interpretability by using the choice probability functions.•Provided... Although researchers increasingly use deep neural networks (DNN) to analyze individual choices, overfitting and interpretability issues remain obstacles in... |
| SourceID | proquest crossref elsevier |
| SourceType | Aggregation Database Enrichment Source Index Database Publisher |
| StartPage | 60 |
| SubjectTerms | Approximation Artificial neural networks Choice learning Choice modeling Deep neural networks Economic forecasting Interpretability Learning Learning theory Logit models Machine learning Mathematical analysis Mathematical models Neural networks Predictions Statistical analysis Statistical learning theory Statistical models Trip surveys Upper bounds |
| Title | Deep neural networks for choice analysis: A statistical learning theory perspective |
| URI | https://dx.doi.org/10.1016/j.trb.2021.03.011 https://www.proquest.com/docview/2544270465 |
| Volume | 148 |
| WOSCitedRecordID | wos000655548800003&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| hasFullText | 1 |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| journalDatabaseRights | – providerCode: PRVESC databaseName: Elsevier SD Freedom Collection Journals 2021 customDbUrl: eissn: 1879-2367 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0003401 issn: 0191-2615 databaseCode: AIEXJ dateStart: 19950201 isFulltext: true titleUrlDefault: https://www.sciencedirect.com providerName: Elsevier |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwtV3db9MwELfKhgR7QDBAbAzkB8QDVarGdr54K9AJUFVAdKjixXISW-1UZaXLpu2_4U_l_JVmRVTwwEsUOXES-X45n893v0PohcgSmYtQBCWYnwFjSRZkJKYBQKWMaEmVMCH_30bJeJxOp9nnTuenz4W5XCRVlV5dZcv_KmpoA2Hr1Nl_EHfzUGiAcxA6HEHscPwrwb-TctnVNJUw-JUN8jacC11QdHOTHmBpSGxOuk4oMlzNcPfCu0lMduO1pjRuJ2J6G7bhQ7fYcXRBsx5Yo6u6-0Y_19al9np17bR3vumZrGbibLP5C7z6er52rc59XWxRN-j7Dv0M8ObV7EK0XRakFVpl_Wg-l8arr3UAk3FxZmEA6zq71y2tWk4BQ5pr7obethSdTvPaqgS_TQjWN3Haq1d5T3-KYbR16v0G-fb4Ez8-GY34ZDidvFz-CHRdMr1_74q03EK7JIky0Ju7gw_D6cdmtqes72pe2q_2O-cmhnDjrX-yfTasAGPaTO6je25NggcWSw9QR1b76I5PWT_fR3st1sqH6KtGGLYIwx5hGBCGLcKwR9hrPMAtfGGPL2zxhVv4eoROjoeTt-8DV5sjKGiU1gEtaMmIEH0pRR6yiIqwSIhQpcwzJrMwETATkCKNJFOsVHFMipLIMM7BAFaJUvQx2qnOKvkEYVlEJC6FohGTLGcxrEiSuIDJRuVpqMr4APX9qPHCEdfr-ikL7iMUTzkMNNcDzfuUw0AfoFdNl6Vlbdl2M_Oi4M7stOYkBxBt63bkxcbd73_ONeEfSfosjg63X36K7q7_jCO0U68u5DN0u7gEmayeO5D9AnHUrUQ |
| linkProvider | Elsevier |
| openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Deep+neural+networks+for+choice+analysis%3A+A+statistical+learning+theory+perspective&rft.jtitle=Transportation+research.+Part+B%3A+methodological&rft.au=Wang%2C+Shenhao&rft.au=Wang%2C+Qingyi&rft.au=Bailey%2C+Nate&rft.au=Zhao%2C+Jinhua&rft.date=2021-06-01&rft.pub=Elsevier+Science+Ltd&rft.issn=0191-2615&rft.eissn=1879-2367&rft.volume=148&rft.spage=60&rft_id=info:doi/10.1016%2Fj.trb.2021.03.011&rft.externalDBID=NO_FULL_TEXT |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0191-2615&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0191-2615&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0191-2615&client=summon |