{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,2,21]],"date-time":"2025-02-21T07:20:47Z","timestamp":1740122447405,"version":"3.37.3"},"reference-count":37,"publisher":"Springer Science and Business Media LLC","issue":"2","license":[{"start":{"date-parts":[[2023,1,23]],"date-time":"2023-01-23T00:00:00Z","timestamp":1674432000000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2023,1,23]],"date-time":"2023-01-23T00:00:00Z","timestamp":1674432000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"DOI":"10.13039\/501100003406","name":"Tekes","doi-asserted-by":"publisher","award":["6878"],"id":[{"id":"10.13039\/501100003406","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Data Min Knowl Disc"],"published-print":{"date-parts":[[2023,3]]},"abstract":"Abstract<\/jats:title>Uplift modeling refers to individual level causal inference. Existing research on the topic ignores one prevalent and important aspect: high class imbalance. For instance in online environments uplift modeling is used to optimally target ads and discounts, but very few users ever end up clicking an ad or buying. One common approach to deal with imbalance in classification is by undersampling the dataset. In this work, we show how undersampling can be extended to uplift modeling. We propose four undersampling methods for uplift modeling. We compare the proposed methods empirically and show when some methods have a tendency to break down. One key observation is that accounting for the imbalance is particularly important for uplift random forests, which explains the poor performance of the model in earlier works. Undersampling is also crucial for class-variable transformation based models.<\/jats:p>","DOI":"10.1007\/s10618-023-00917-9","type":"journal-article","created":{"date-parts":[[2023,1,23]],"date-time":"2023-01-23T08:02:49Z","timestamp":1674460969000},"page":"736-766","update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":4,"title":["Exploring uplift modeling with high class imbalance"],"prefix":"10.1007","volume":"37","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-8481-9325","authenticated-orcid":false,"given":"Otto","family":"Nyberg","sequence":"first","affiliation":[]},{"given":"Arto","family":"Klami","sequence":"additional","affiliation":[]}],"member":"297","published-online":{"date-parts":[[2023,1,23]]},"reference":[{"key":"917_CR1","unstructured":"Athey S, Imbens G (2015) Recursive partitioning for heterogeneous causal effects. arXiv arXiv:1504.01132"},{"issue":"3","key":"917_CR2","doi-asserted-by":"publisher","first-page":"399","DOI":"10.1080\/00273171.2011.568786","volume":"46","author":"PC Austin","year":"2011","unstructured":"Austin PC (2011) An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivar Behav Res 46(3):399\u2013424","journal-title":"Multivar Behav Res"},{"key":"917_CR3","unstructured":"Belbahri M, Gandouet O, Kazma G (2020) Adapting neural networks for uplift models. arXiv:2011.00041"},{"key":"917_CR4","unstructured":"Belbahri M, Gandouet O, Murua A et al (2021) A twin neural model for uplift. arxiv:2105.05146"},{"key":"917_CR5","first-page":"47","volume-title":"Lecture notes in computer science","author":"A Betlei","year":"2018","unstructured":"Betlei A, Diemert E, Amini MR (2018) Uplift prediction with dependent feature representation in imbalanced treatment and control conditions. Lecture notes in computer science, vol V. Springer, Cham, pp 47\u201355"},{"key":"917_CR6","doi-asserted-by":"publisher","first-page":"321","DOI":"10.1613\/jair.953","volume":"16","author":"NV Chawla","year":"2002","unstructured":"Chawla NV, Bowyer KW, Hall LO et al (2002) Smote: synthetic minority over-sampling technique. J Artif Intell Res 16:321\u2013357","journal-title":"J Artif Intell Res"},{"key":"917_CR7","doi-asserted-by":"publisher","first-page":"497","DOI":"10.1016\/j.ins.2019.12.075","volume":"548","author":"F Devriendt","year":"2021","unstructured":"Devriendt F, Berrevoets J, Verbeke W (2021) Why you should stop predicting customer churn and start using uplift models. Inf Sci 548:497\u2013515","journal-title":"Inf Sci"},{"key":"917_CR8","unstructured":"Diemert E, Betlei A, Renaudin C et al (2018) A large scale benchmark for uplift modeling. In: Proceedings of the AdKDD and TargetAd workshop, KDD, London, United Kingdom, August, 20, 2018"},{"key":"917_CR9","first-page":"1","volume":"23","author":"C Fern\u00e1ndez-Lor\u00eda","year":"2022","unstructured":"Fern\u00e1ndez-Lor\u00eda C, Provost F (2022) Causal classification: treatment effect vs. outcome prediction. J Mach Learn Res 23:1\u201335","journal-title":"J Mach Learn Res"},{"issue":"2","key":"917_CR10","doi-asserted-by":"publisher","first-page":"647","DOI":"10.1016\/j.ejor.2019.11.030","volume":"283","author":"RM Gubela","year":"2020","unstructured":"Gubela RM, Lessmann S, Jaroszewicz S (2020) Response transformation and profit decomposition for revenue uplift modeling. Eur J Oper Res 283(2):647\u2013661","journal-title":"Eur J Oper Res"},{"issue":"3\u20134","key":"917_CR11","doi-asserted-by":"publisher","first-page":"230","DOI":"10.1080\/01969722.2015.1012892","volume":"46","author":"L Guelman","year":"2015","unstructured":"Guelman L, Guill\u00e9n M, P\u00e9rez-Mar\u00edn AM (2015) Uplift random forests. Cybern Syst 46(3\u20134):230\u2013248","journal-title":"Cybern Syst"},{"key":"917_CR12","unstructured":"Gutierrez P, G\u00e9rardy JY (2017) Causal inference and uplift modelling: a review of the literature. In: Proceedings of the 3rd international conference on predictive applications and APIs, vol 67, pp 1\u201313"},{"key":"917_CR13","doi-asserted-by":"publisher","unstructured":"Haupt J, Lessmann S (2020) Targeting customers under response-dependent costs. arxiv:2003.06271. https:\/\/doi.org\/10.1016\/j.ejor.2021.05.045","DOI":"10.1016\/j.ejor.2021.05.045"},{"key":"917_CR14","unstructured":"Jaskowski M, Jaroszewicz S (2012) Uplift modeling for clinical trial data. ICML workshop on clinical data analysis"},{"key":"917_CR15","unstructured":"Johansson FD, Shalit U, Sontag D (2016) Learning representations for counterfactual inference. In: Proceedings of the 33rd international conference on machine learning"},{"issue":"4","key":"917_CR16","first-page":"1","volume":"52","author":"H Kaur","year":"2019","unstructured":"Kaur H, Pannu HS, Malhi AK (2019) A systematic review on imbalanced data challenges in machine learning: applications and solutions. ACM Comput Surv 52(4):1\u201336","journal-title":"ACM Comput Surv"},{"key":"917_CR17","doi-asserted-by":"publisher","first-page":"4156","DOI":"10.1073\/pnas.1804597116","volume":"116","author":"SR K\u00fcnzel","year":"2019","unstructured":"K\u00fcnzel SR, Sekhon JS, Bickel PJ et al (2019) Metalearners for estimating heterogeneous treatment effects using machine learning. Proc Natl Acad Sci USA 116:4156\u20134165","journal-title":"Proc Natl Acad Sci USA"},{"key":"917_CR18","unstructured":"Lai LYT (2006) Influential marketing: a new direct marketing strategy addressing the existence of voluntary buyers. Ph.D. thesis, University of British Columbia"},{"key":"917_CR19","doi-asserted-by":"publisher","first-page":"78","DOI":"10.1145\/772862.772872","volume":"4","author":"VS Lo","year":"2002","unstructured":"Lo VS (2002) The true lift model\u2014a novel data mining approach to response modeling in database marketing. SIGKDD Explor 4:78\u201386","journal-title":"SIGKDD Explor"},{"key":"917_CR20","unstructured":"Naeini MP, Cooper GF, Hauskrecht M (2015) Obtaining well calibrated probabilities using Bayesian binning. In: Proceedings of the twenty-ninth AAAI conference on artificial intelligence, vol 2015, pp 2901\u20132907"},{"key":"917_CR21","doi-asserted-by":"crossref","unstructured":"Nyberg O, Ku\u015bmierczyk T, Klami A (2021) Uplift modeling with high class imbalance. In: Proceedings of the 13th Asian conference on machine learning, pp 315\u2013330","DOI":"10.21203\/rs.3.rs-1719505\/v1"},{"issue":"2","key":"917_CR22","doi-asserted-by":"publisher","first-page":"273","DOI":"10.1007\/s10618-019-00670-y","volume":"34","author":"D Olaya","year":"2020","unstructured":"Olaya D, Coussement K, Verbeke W (2020) A survey and benchmarking study of multitreatment uplift modeling. Data Min Knowl Disc 34(2):273\u2013308","journal-title":"Data Min Knowl Disc"},{"key":"917_CR23","unstructured":"Papangelou K (2021) Assessing treatment effect heterogeneity: predictive covariate selection and subgroup identification. Ph.D. thesis, University of Manchester"},{"key":"917_CR24","doi-asserted-by":"publisher","first-page":"96","DOI":"10.1214\/09-SS057","volume":"3","author":"J Pearl","year":"2009","unstructured":"Pearl J (2009) Causal inference in statistics: an overview. Stat Surv 3:96\u2013146","journal-title":"Stat Surv"},{"key":"917_CR25","first-page":"2825","volume":"12","author":"F Pedregosa","year":"2011","unstructured":"Pedregosa F, Varoquaux G, Gramfort A et al (2011) Scikit-learn: machine learning in python. J Mach Learn Res 12:2825\u20132830","journal-title":"J Mach Learn Res"},{"key":"917_CR26","first-page":"61","volume":"10","author":"J Platt","year":"1999","unstructured":"Platt J (1999) Probabilistic outputs for support vector machines and comparisons to regularized likelihood methods. Adv Large Margin Classif 10:61\u201374","journal-title":"Adv Large Margin Classif"},{"key":"917_CR27","unstructured":"Radcliffe NJ (2008) Hillstrom\u2019s MineThatData email analytics challenge: an approach using uplift modelling. Response, pp 1\u201319. http:\/\/stochasticsolutions.com\/"},{"key":"917_CR28","unstructured":"Radcliffe NJ, Surry PD (1999) Differential response analysis: modelling true response by isolating the effect of a single action. Credit scoring and credit control VI"},{"key":"917_CR29","doi-asserted-by":"crossref","unstructured":"Richardson M, Ragno R, Dominowska E (2007) Predicting clicks: estimating the click-through rate for new ads. In: Proceedings of the 16th international conference on World Wide Web, pp 521\u2013529","DOI":"10.1145\/1242572.1242643"},{"key":"917_CR30","doi-asserted-by":"crossref","unstructured":"R\u00f6ssler J, Tilly R, Schoder D (2021) To treat, or not to treat: reducing volatility in uplift modeling through weighted ensembles. In: Proceedings of the 54th Hawaii international conference on system sciences","DOI":"10.24251\/HICSS.2021.193"},{"key":"917_CR31","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1007\/s10618-018-0576-8","volume":"32","author":"K Ruda\u015b","year":"2018","unstructured":"Ruda\u015b K, Jaroszewicz S (2018) Linear regression for uplift modeling. Data Min Knowl Discov 32:1\u201331","journal-title":"Data Min Knowl Discov"},{"key":"917_CR32","doi-asserted-by":"crossref","unstructured":"Rzepakowski P, Jaroszewicz S (2010) Decision trees for uplift modeling. In: Proceedings\u2014IEEE international conference on data mining, ICDM pp 441\u2013450","DOI":"10.1109\/ICDM.2010.62"},{"key":"917_CR33","unstructured":"Semenova D, Temirkaeva M (2019) The comparison of methods for individual treatment effect detection. In: CEUR workshop proceedings, pp 46\u201356"},{"key":"917_CR34","doi-asserted-by":"publisher","first-page":"211","DOI":"10.1016\/j.ejor.2011.09.031","volume":"218","author":"W Verbeke","year":"2012","unstructured":"Verbeke W, Dejaeger K, Martens D et al (2012) New insights into churn prediction in the telecommunication sector: a profit driven data mining approach. Eur J Oper Res 218:211\u2013229","journal-title":"Eur J Oper Res"},{"issue":"523","key":"917_CR35","doi-asserted-by":"publisher","first-page":"1228","DOI":"10.1080\/01621459.2017.1319839","volume":"113","author":"S Wager","year":"2018","unstructured":"Wager S, Athey S (2018) Estimation and inference of heterogeneous treatment effects using random forests. J Am Stat Assoc 113(523):1228\u20131242","journal-title":"J Am Stat Assoc"},{"key":"917_CR36","doi-asserted-by":"crossref","unstructured":"Zadrozny B, Elkan C (2002) Transforming classifier scores into accurate multiclass probability estimates. In: Proceedings of the eighth ACM SIGKDD international conference on Knowledge discovery and data mining KDD, vol 02, pp 694\u2013699","DOI":"10.1145\/775047.775151"},{"key":"917_CR37","doi-asserted-by":"crossref","unstructured":"Zhao Z, Zhang Y, Harinen T, et\u00a0al (2022) Feature selection methods for uplift modeling and heterogeneous treatment effect. In: IFIP advances in information and communication technology, pp 217\u2013230","DOI":"10.1007\/978-3-031-08337-2_19"}],"container-title":["Data Mining and Knowledge Discovery"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10618-023-00917-9.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s10618-023-00917-9\/fulltext.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10618-023-00917-9.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,2,24]],"date-time":"2023-02-24T11:35:48Z","timestamp":1677238548000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1007\/s10618-023-00917-9"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,1,23]]},"references-count":37,"journal-issue":{"issue":"2","published-print":{"date-parts":[[2023,3]]}},"alternative-id":["917"],"URL":"https:\/\/doi.org\/10.1007\/s10618-023-00917-9","relation":{"has-preprint":[{"id-type":"doi","id":"10.21203\/rs.3.rs-1719505\/v1","asserted-by":"object"}]},"ISSN":["1384-5810","1573-756X"],"issn-type":[{"type":"print","value":"1384-5810"},{"type":"electronic","value":"1573-756X"}],"subject":[],"published":{"date-parts":[[2023,1,23]]},"assertion":[{"value":"2 June 2022","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"2 January 2023","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"23 January 2023","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"Neither author has any conflicts of interest to disclose.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Conflict of interest"}}]}}