×

FLAME: a fast large-scale almost matching exactly approach to causal inference. (English) Zbl 07370548

Summary: A classical problem in causal inference is that of matching, where treatment units need to be matched to control units based on covariate information. In this work, we propose a method that computes high quality almost-exact matches for high-dimensional categorical datasets. This method, called FLAME (Fast Large-scale Almost Matching Exactly), learns a distance metric for matching using a hold-out training data set. In order to perform matching efficiently for large datasets, FLAME leverages techniques that are natural for query processing in the area of database management, and two implementations of FLAME are provided: the first uses SQL queries and the second uses bit-vector techniques. The algorithm starts by constructing matches of the highest quality (exact matches on all covariates), and successively eliminates variables in order to match exactly on as many variables as possible, while still maintaining interpretable high-quality matches and balance between treatment and control groups. We leverage these high quality matches to estimate conditional average treatment effects (CATEs). Our experiments show that FLAME scales to huge datasets with millions of observations where existing state-of-the-art methods fail, and that it achieves significantly better performance than other matching methods.

MSC:

68T05 Learning and adaptive systems in artificial intelligence
PDFBibTeX XMLCite
Full Text: arXiv Link

References:

[1] Alberto Abadie and Guido W Imbens. A martingale representation for matching estimators. Journal of the American Statistical Association, 107(498):833-843, 2012. · Zbl 1261.62008
[2] Jason Abrevaya. Estimating the effect of smoking on birth outcomes using a matched panel data approach.Journal of Applied Econometrics, 21(4):489-519, 2006.
[3] Kathleen E Adams, Vincent P Miller, Carla Ernst, Brenda K Nishimura, Cathy Melvin, and Robert Merritt. Neonatal health care costs related to smoking during pregnancy. Health Economics, 11(3):193-206, 2002.
[4] Susan Athey, Julie Tibshirani, Stefan Wager, et al. Generalized random forests.The Annals of Statistics, 47(2):1148-1178, 2019. · Zbl 1418.62102
[5] M. Usaid Awan, Yameng Liu, Marco Morucci, Sudeepa Roy, Cynthia Rudin, and Alexander Volfovsky. Interpretable almost matching exactly with instrumental variables. In Proceedings of the Thirty-Fifth Conference on Uncertainty in Artificial Intelligence, UAI, 2019.
[6] M. Usaid Awan, Marco Morucci, Vittorio Orlandi, Sudeepa Roy, Cynthia Rudin, and Alexander Volfovsky.Almost-matching-exactly for treatment effect estimation under network interference.Proceedings of The 23rd International Conference on Artificial Intelligence and Statistics, AISTATS, 2020.
[7] Alexandre Belloni, Victor Chernozhukov, and Christian Hansen. Inference on treatment effects after selection among high-dimensional controls.The Review of Economic Studies, 81(2):608-650, 2014. · Zbl 1409.62142
[8] M Alan Brookhart, Sebastian Schneeweiss, Kenneth J Rothman, Robert J Glynn, Jerry Avorn, and Til St¨urmer. Variable selection for propensity score models.American Journal of Epidemiology, 163(12):1149-1156, 2006.
[9] Matias D Cattaneo and Max H Farrell. Efficient estimation of the dose-response function under ignorability using subclassification on the covariates.Advances in Econometrics, 27:93, 2011. · Zbl 1443.62027
[10] F.S. Chapin.Experimental Designs in Sociological Research. Harper; New York, 1947.
[11] Hugh A Chipman, Edward I George, Robert E McCulloch, et al. BART: Bayesian additive regression trees.The Annals of Applied Statistics, 4(1):266-298, 2010. · Zbl 1189.62066
[12] William G Cochran and Donald B Rubin. Controlling bias in observational studies: A review.Sankhy¯a: The Indian Journal of Statistics, Series A, pages 417-446, 1973. · Zbl 0291.62012
[13] Alexis Diamond and Jasjeet S. Sekhon. Genetic matching for estimating causal effects: A general multivariate matching method for achieving balance in observational studies.The Review of Economics and Statistics, 95(3):932-945, 2013.
[14] Awa Dieng, Yameng Liu, Sudeepa Roy, Cynthia Rudin, and Alexander Volfovsky. Interpretable almost-exact matching for causal inference. InProceedings of Artificial Intelligence and Statistics (AISTATS), pages 2445-2453, 2019.
[15] Vincent Dorie, Jennifer Hill, Uri Shalit, Marc Scott, and Dan Cervone. Automated versus do-it-yourself methods for causal inference: Lessons learned from a data analysis competition.Statistical Science, 34(1):43-68, 2019. · Zbl 1420.62345
[16] MA Efroymson. Multiple regression analysis.Mathematical Methods for Digital Computers, pages 191-203, 1960.
[17] Max H Farrell. Robust inference on average treatment effects with possibly more covariates than observations.Journal of Econometrics, 189(1):1-23, 2015. · Zbl 1337.62113
[18] Max H Farrell, Tengyuan Liang, and Sanjog Misra. Deep neural networks for estimation and inference: Application to causal effects and other semiparametric estimands.arXiv preprint arXiv:1809.09953, 2018.
[19] Ernest Greenwood.Experimental Sociology: A Study in Method. King’s Crown Press, 1945.
[20] Neha R. Gupta, Vittorio Orlandi, Chia-Rui Chang, Tianyu Wang, Marco Morucci, Pritam Dey, Thomas J. Howell, Xian Sun, Angikar Ghosal, Sudeepa Roy, Cynthia Rudin, and Alexander Volfovsky. dame-flame: A python library providing fast interpretable matching for causal inference. arXiv 2101.01867, 2021.
[21] Jinyong Hahn. Functional restriction and efficiency in causal inference.Review of Economics and Statistics, 86(1):73-76, 2004.
[22] G¨unter J Hitsch and Sanjog Misra. Heterogeneous treatment effects and optimal targeting policy evaluation.Available at SSRN 3111957, 2018.
[23] MA Honein, LJ Paulozzi, and ML Watkins. Maternal smoking and birth defects: validity of birth certificate data for effect estimation.Public Health Reports, 116(4):327, 2001.
[24] Stefano M Iacus, Gary King, and Giuseppe Porro. Causal inference without balance checking: Coarsened exact matching.Political Analysis, 20(1):1-24, 2011a. · Zbl 1396.62011
[25] Stefano M Iacus, Gary King, and Giuseppe Porro. Multivariate matching methods that are monotonic imbalance bounding.Journal of the American Statistical Association, 106 (493):345-361, 2011b. · Zbl 1396.62011
[26] Guido W Imbens and Donald B Rubin.Causal inference in statistics, social, and biomedical sciences. Cambridge University Press, 2015. · Zbl 1355.62002
[27] Mariana Caricati Kataoka, Ana Paula Pinho Carvalheira, Anna Paula Ferrari, Ma´ıra Barreto Malta, Maria Antonieta de Barros Leite Carvalhaes, and Cristina Maria Garcia de Lima Parada. Smoking during pregnancy and harm reduction in birth weight: a cross-sectional study.BMC Pregnancy and Childbirth, 18(1):1-10, 2018.
[28] Anthony J Kondracki. Prevalence and patterns of cigarette smoking before and during early and late pregnancy according to maternal characteristics: the first national data based on the 2003 birth certificate revision, United States, 2016.Reproductive Health, 16(1): 142, 2019.
[29] Anthony J Kondracki. Low birthweight in term singletons mediates the association between maternal smoking intensity exposure status and immediate neonatal intensive care unit admission: the e-value assessment.BMC Pregnancy and Childbirth, 20:1-9, 2020.
[30] M. Lichman. UCI machine learning repository, 2013. URLhttp://archive.ics.uci.edu/ ml.
[31] Malini Mahendra, Martina Steurer-Muller, Samuel F Hohmann, Roberta L Keller, Anil Aswani, and R Adams Dudley.Predicting NICU admissions in near-term and term infants with low illness acuity.Journal of Perinatology, pages 1-8, 2020.
[32] Robert L McCornack. A comparison of three predictor selection techniques in multiple regression.Psychometrika, 35(2):257-271, 1970. · Zbl 0195.48703
[33] Marco Morucci, Md. Noor-E-Alam, and Cynthia Rudin. Hypothesis tests that are robust to choice of matching method.arXiv preprint arXiv:1812.02227, 2018.
[34] Marco Morucci, Vittorio Orlandi, Sudeepa Roy, Cynthia Rudin, and Alexander Volfovsky. Adaptive hyper-box matching for interpretable individualized treatment effect estimation. Proceedings of the Thirty-Sixth Conference on Uncertainty in Artificial Intelligence, UAI, 2020.
[35] National Center for Health Statistics, NCHS. User guide to the 2010 natality public use file. Technical report, Centers for Disease Control and Prevention (CDC), 2010.
[36] M. Noor-E-Alam and C. Rudin. Robust nonparametric testing for causal inference in natural experiments.Working paper, 2015.
[37] Harsh Parikh, Cynthia Rudin, and Alexander Volfovsky. MALTS: Matching After Learning to Stretch.arXiv e-prints: arXiv:1811.07415, Nov 2018.
[38] Harsh Parikh, Cynthia Rudin, and Alexander Volfovsky. An application of matching after learning to stretch (MALTS) to the ACIC 2018 causal inference challenge data.Observational Studies, pages 118-130, 2019.
[39] PostgreSQL.PostgreSQL, 2016. URLhttp://www.postgresql.org.
[40] Jeremy A Rassen and Sebastian Schneeweiss. Using high-dimensional propensity scores to automate confounding control in a distributed medical product safety surveillance system. Pharmacoepidemiology and drug safety, 21(S1):41-49, 2012.
[41] Mar´ıa Resa and Jos´e R Zubizarreta. Evaluation of subset matching methods and forms of covariate balance.Statistics in Medicine, 35(27):4961-4979, 2016.
[42] Paul R Rosenbaum. Imposing minimax and quantile constraints on optimal matching in observational studies.Journal of Computational and Graphical Statistics, 2016.
[43] Paul R Rosenbaum and Donald B Rubin.The central role of the propensity score in observational studies for causal effects.Biometrika, 70(1):41-55, 1983. · Zbl 0522.62091
[44] Michelle E Ross, Amanda R Kreider, Yuan-Shung Huang, Meredith Matone, David M Rubin, and A Russell Localio. Propensity score methods for analyzing observational data like randomized experiments: challenges and solutions for rare outcomes and exposures. American Journal of Epidemiology, 181(12):989-995, 2015.
[45] Donald B Rubin. Matching to remove bias in observational studies.Biometrics, pages 159-183, 1973a.
[46] Donald B Rubin. The use of matched sampling and regression adjustment to remove bias in observational studies.Biometrics, pages 185-203, 1973b.
[47] Donald B Rubin. Multivariate matching methods that are equal percent bias reducing, i: Some examples.Biometrics, pages 109-120, 1976. · Zbl 0326.62043
[48] Donald B. Rubin. Causal inference using potential outcomes: Design, modeling, decisions. Journal of the American Statistical Association, 100:322-331, 2005. · Zbl 1117.62418
[49] Cynthia Rudin. Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead.Nature Machine Intelligence, 1:206-215, May 2019.
[50] Sebastian Schneeweiss, Jeremy A Rassen, Robert J Glynn, Jerry Avorn, Helen Mogun, and M Alan Brookhart. High-dimensional propensity score adjustment in studies of treatment effects using health care claims data.Epidemiology, 20(4):512, 2009.
[51] Mark J Van Der Laan and Daniel Rubin. Targeted maximum likelihood learning.The International Journal of Biostatistics, 2(1), 2006.
[52] Stefan Wager and Susan Athey. Estimation and inference of heterogeneous treatment effects using random forests.Journal of the American Statistical Association, 113(523):1228- 1242, 2018. · Zbl 1402.62056
[53] Jay L Zagorsky. Marriage and divorce’s impact on wealth.Journal of Sociology, 41(4): 406-424, 2005.
[54] Jos´e R Zubizarreta. Using mixed integer programming for matching in an observational study of kidney failure after surgery.Journal of the American Statistical Association, 107 (500):1360-1371, 2012. · Zbl 1258.62119
[55] Jos´e R Zubizarreta, Ricardo D Paredes, Paul R Rosenbaum, et al. Matching for balance, pairing for heterogeneity in an observational study of the effectiveness of for-profit and not-for-profit high schools in chile.The Annals of Applied Statistics, 8(1):204-231, 2014. · Zbl 1454.62510
[56] Jos´e R. Zubizarreta and Luke Keele. Optimal multilevel matching in clustered observational studies: A case study of the effectiveness of private schools under a large-scale voucher system.Journal of the American Statistical Association, 112(518), 2017.
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.