Have it both ways : from A/B testing to A&B testing with exceptional model mining

Wouter Duivesteijn, Tara Farzami, Thijs Putman, Evertjan Peer, Hilde J.P. Weerts, Jasper N. Adegeest, Gerson Foks, Mykola Pechenizkiy

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

1 Citation (Scopus)
1 Downloads (Pure)

Abstract

In traditional A/B testing, we have two variants of the same product, a pool of test subjects, and a measure of success. In a randomized experiment, each test subject is presented with one of the two variants, and the measure of success is aggregated per variant. The variant of the product associated with the most success is retained, while the other variant is discarded. This, however, presumes that the company producing the products only has enough capacity to maintain one of the two product variants. If more capacity is available, then advanced data science techniques can extract more profit for the company from the A/B testing results. Exceptional Model Mining is one such advanced data science technique, which specializes in identifying subgroups that behave differently from the overall population. Using the association model class for EMM, we can find subpopulations that prefer variant A where the general population prefers variant B, and vice versa. This data science technique is applied on data from StudyPortals, a global study choice platform that ran an A/B test on the design of aspects of their website.

Original languageEnglish
Title of host publicationMachine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2017, Proceedings
EditorsMichelangelo Ceci, Saso Dzeroski, Donato Malerba, Yasemin Altun, Kamalika Das, Jesse Read, Marinka Zitnik, Jerzy Stefanowski, Taneli Mielikäinen
Place of PublicationCham
PublisherSpringer
Pages114-126
Number of pages13
ISBN (Electronic)978-3-319-71273-4
ISBN (Print)978-3-319-71272-7
DOIs
Publication statusPublished - 30 Dec 2017
EventEuropean Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2017 - Skopje, Macedonia, The Former Yugoslav Republic of
Duration: 18 Sep 201722 Sep 2017

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10536 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

ConferenceEuropean Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2017
CountryMacedonia, The Former Yugoslav Republic of
CitySkopje
Period18/09/1722/09/17

Fingerprint

Mining
Testing
Websites
Industry
Profitability
Randomized Experiments
Association Model
Model
Profit
Experiments
Subgroup

Keywords

  • A/B testing
  • Association
  • E-commerce
  • Exceptional Model Mining
  • Online controlled experiments
  • Website optimization

Cite this

Duivesteijn, W., Farzami, T., Putman, T., Peer, E., Weerts, H. J. P., Adegeest, J. N., ... Pechenizkiy, M. (2017). Have it both ways : from A/B testing to A&B testing with exceptional model mining. In M. Ceci, S. Dzeroski, D. Malerba, Y. Altun, K. Das, J. Read, M. Zitnik, J. Stefanowski, ... T. Mielikäinen (Eds.), Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2017, Proceedings (pp. 114-126). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10536 LNAI). Cham : Springer. https://doi.org/10.1007/978-3-319-71273-4_10
Duivesteijn, Wouter ; Farzami, Tara ; Putman, Thijs ; Peer, Evertjan ; Weerts, Hilde J.P. ; Adegeest, Jasper N. ; Foks, Gerson ; Pechenizkiy, Mykola. / Have it both ways : from A/B testing to A&B testing with exceptional model mining. Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2017, Proceedings. editor / Michelangelo Ceci ; Saso Dzeroski ; Donato Malerba ; Yasemin Altun ; Kamalika Das ; Jesse Read ; Marinka Zitnik ; Jerzy Stefanowski ; Taneli Mielikäinen. Cham : Springer, 2017. pp. 114-126 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{1ec96bb22a5148b4ae53c80544f160a0,
title = "Have it both ways : from A/B testing to A&B testing with exceptional model mining",
abstract = "In traditional A/B testing, we have two variants of the same product, a pool of test subjects, and a measure of success. In a randomized experiment, each test subject is presented with one of the two variants, and the measure of success is aggregated per variant. The variant of the product associated with the most success is retained, while the other variant is discarded. This, however, presumes that the company producing the products only has enough capacity to maintain one of the two product variants. If more capacity is available, then advanced data science techniques can extract more profit for the company from the A/B testing results. Exceptional Model Mining is one such advanced data science technique, which specializes in identifying subgroups that behave differently from the overall population. Using the association model class for EMM, we can find subpopulations that prefer variant A where the general population prefers variant B, and vice versa. This data science technique is applied on data from StudyPortals, a global study choice platform that ran an A/B test on the design of aspects of their website.",
keywords = "A/B testing, Association, E-commerce, Exceptional Model Mining, Online controlled experiments, Website optimization",
author = "Wouter Duivesteijn and Tara Farzami and Thijs Putman and Evertjan Peer and Weerts, {Hilde J.P.} and Adegeest, {Jasper N.} and Gerson Foks and Mykola Pechenizkiy",
year = "2017",
month = "12",
day = "30",
doi = "10.1007/978-3-319-71273-4_10",
language = "English",
isbn = "978-3-319-71272-7",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer",
pages = "114--126",
editor = "Michelangelo Ceci and Saso Dzeroski and Donato Malerba and Yasemin Altun and Kamalika Das and Jesse Read and Marinka Zitnik and Jerzy Stefanowski and Taneli Mielik{\"a}inen",
booktitle = "Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2017, Proceedings",
address = "Germany",

}

Duivesteijn, W, Farzami, T, Putman, T, Peer, E, Weerts, HJP, Adegeest, JN, Foks, G & Pechenizkiy, M 2017, Have it both ways : from A/B testing to A&B testing with exceptional model mining. in M Ceci, S Dzeroski, D Malerba, Y Altun, K Das, J Read, M Zitnik, J Stefanowski & T Mielikäinen (eds), Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2017, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10536 LNAI, Springer, Cham , pp. 114-126, European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2017, Skopje, Macedonia, The Former Yugoslav Republic of, 18/09/17. https://doi.org/10.1007/978-3-319-71273-4_10

Have it both ways : from A/B testing to A&B testing with exceptional model mining. / Duivesteijn, Wouter; Farzami, Tara; Putman, Thijs; Peer, Evertjan; Weerts, Hilde J.P.; Adegeest, Jasper N.; Foks, Gerson; Pechenizkiy, Mykola.

Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2017, Proceedings. ed. / Michelangelo Ceci; Saso Dzeroski; Donato Malerba; Yasemin Altun; Kamalika Das; Jesse Read; Marinka Zitnik; Jerzy Stefanowski; Taneli Mielikäinen. Cham : Springer, 2017. p. 114-126 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10536 LNAI).

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

TY - GEN

T1 - Have it both ways : from A/B testing to A&B testing with exceptional model mining

AU - Duivesteijn, Wouter

AU - Farzami, Tara

AU - Putman, Thijs

AU - Peer, Evertjan

AU - Weerts, Hilde J.P.

AU - Adegeest, Jasper N.

AU - Foks, Gerson

AU - Pechenizkiy, Mykola

PY - 2017/12/30

Y1 - 2017/12/30

N2 - In traditional A/B testing, we have two variants of the same product, a pool of test subjects, and a measure of success. In a randomized experiment, each test subject is presented with one of the two variants, and the measure of success is aggregated per variant. The variant of the product associated with the most success is retained, while the other variant is discarded. This, however, presumes that the company producing the products only has enough capacity to maintain one of the two product variants. If more capacity is available, then advanced data science techniques can extract more profit for the company from the A/B testing results. Exceptional Model Mining is one such advanced data science technique, which specializes in identifying subgroups that behave differently from the overall population. Using the association model class for EMM, we can find subpopulations that prefer variant A where the general population prefers variant B, and vice versa. This data science technique is applied on data from StudyPortals, a global study choice platform that ran an A/B test on the design of aspects of their website.

AB - In traditional A/B testing, we have two variants of the same product, a pool of test subjects, and a measure of success. In a randomized experiment, each test subject is presented with one of the two variants, and the measure of success is aggregated per variant. The variant of the product associated with the most success is retained, while the other variant is discarded. This, however, presumes that the company producing the products only has enough capacity to maintain one of the two product variants. If more capacity is available, then advanced data science techniques can extract more profit for the company from the A/B testing results. Exceptional Model Mining is one such advanced data science technique, which specializes in identifying subgroups that behave differently from the overall population. Using the association model class for EMM, we can find subpopulations that prefer variant A where the general population prefers variant B, and vice versa. This data science technique is applied on data from StudyPortals, a global study choice platform that ran an A/B test on the design of aspects of their website.

KW - A/B testing

KW - Association

KW - E-commerce

KW - Exceptional Model Mining

KW - Online controlled experiments

KW - Website optimization

UR - http://www.scopus.com/inward/record.url?scp=85040256149&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-71273-4_10

DO - 10.1007/978-3-319-71273-4_10

M3 - Conference contribution

AN - SCOPUS:85040256149

SN - 978-3-319-71272-7

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

SP - 114

EP - 126

BT - Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2017, Proceedings

A2 - Ceci, Michelangelo

A2 - Dzeroski, Saso

A2 - Malerba, Donato

A2 - Altun, Yasemin

A2 - Das, Kamalika

A2 - Read, Jesse

A2 - Zitnik, Marinka

A2 - Stefanowski, Jerzy

A2 - Mielikäinen, Taneli

PB - Springer

CY - Cham

ER -

Duivesteijn W, Farzami T, Putman T, Peer E, Weerts HJP, Adegeest JN et al. Have it both ways : from A/B testing to A&B testing with exceptional model mining. In Ceci M, Dzeroski S, Malerba D, Altun Y, Das K, Read J, Zitnik M, Stefanowski J, Mielikäinen T, editors, Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2017, Proceedings. Cham : Springer. 2017. p. 114-126. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-71273-4_10