Interpretable Almost-Exact Matching for Causal Inference

By Awa Dieng, Yameng Liu, Sudeepa Roy, Cynthia Rudin, Alexander Volfovsky. 2019

25 citations recorded by the source corpus.

Matching methods are heavily used in the social and health sciences due to their interpretability. We aim to create the highest possible quality of treatment-control matches for categorical data in the potential outcomes framework. The method proposed in this work aims to match units on a weighted Hamming distance, taking into account the relative importance of the covariates; the algorithm aims to match units on as many relevant variables as possible. To do this, the algorithm creates a hierarchy of covariate combinations on which to match (similar to downward closure), in the process solving an optimization problem for each unit in order to construct the optimal matches. The algorithm uses a single dynamic program to solve all of the units’ optimization problems simultaneously. Notable advantages of our method over existing matching procedures are its high-quality interpretable matches, versatility in handling different data distributions that may have irrelevant variables, and ability to handle missing data by matching on as many available covariates as possible.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

FLAME: A Fast Large-scale Almost Matching Exactly Approach to Causal Inference (2017, methodology): “There are several natural extensions to the FLAME procedure: First, the less greedy approach of the Dynamic Almost Matching Exactly (DAME) algorithm (Dieng et al., 2019) extends the work in this paper by providing a solution to the FULL-AME problem, without backwards selection.”
MALTS: Matching After Learning to Stretch (2018, methodology): “(2021), Dieng et al. (2019) where a discrete distance metric is learned by considering the prediction quality of the covariate sets.”
The Secrets of Machine Learning: Ten Things You Wish You Had Known Earlier to be More Effective at Data Analysis (2019, background): “al twins. However, in reality, one cannot usually nd a match on all covariates. Some modern matching methods aim to nd matches on as many relevant covariates as possible using machine learning (e.g., [14]). Once units are assigned to matched groups, an estimated treatment eect would be the dierence ”
Almost-Matching-Exactly for Treatment Effect Estimation under Network Interference (2020, methodology): “Almost-Matching-Exactly (AME) (Wang et al., 2019; Dieng et al., 2019; Awan et al., 2019) provides a framework for the above problem that is explicitly geared towards building interpretable, high-quality matches on discrete covariates, which in our setting are the counts of the treated subgraphs in…”
Adaptive Hyper-box Matching for Interpretable Individualized Treatment Effect Estimation (2020, methodology): “…predictions to create matches; 2. methods within the almost-exact-matching (AEM) framework (FLAME, DAME, and MALTS) (Wang, Morucci, et al. 2017; Dieng et al. 2019; Parikh et al. 2018) that leverage a training set for matching, and 3. the causal forest (CF) framework (Wager and Athey 2018),…”
dame-flame: A Python Library Providing Fast Interpretable Matching for Causal Inference (2021, unknown): “One A further advantage of dame-ﬂame is the higher quality of the matched groups generated by DAME and FLAME relative to propensity score matching, as shown by Dieng et al. (2019).”
Interpretable Almost Matching Exactly With Instrumental Variables (2019, methodology): “We create an Almost-Matching Exactly framework (Wang et al., 2019; Dieng et al., 2019) for the purpose of instrumental variable analysis.”