betaFeedback

WorksMathematics

Dirichlet-Tree Allocation Priors for Bayesian Causal Forests

Abstract

Bayesian Causal Forests (BCF) estimate heterogeneous treatment effects with a separate ensemble of treatment trees. The prior on the leaf values of the treatment trees determines whether sampling variability is attributed to the treatment effect. Independent shrinkage priors exert no control over how a tree allocates variability across its leaves, and their total prior variance grows as leaves are added. We propose the Dirichlet-Tree Bayesian Causal Forest, which jointly regularizes the leaf values of each treatment tree through a Dirichlet-Tree prior on their variance shares, generated by stick-breaking along the tree topology. We characterize the spike and tails of the marginal leaf prior, show that the prior adaptively bounds the number of regions of effect heterogeneity, and establish posterior consistency for the forest. In simulations, the model outperforms BCF when heterogeneity is confined to a few regions and reduces spurious heterogeneity elsewhere, and, unlike independent shrinkage priors, its accuracy does not deteriorate as the trees grow. We also evaluate the model on real data, including a pancreatic cancer survival study.

Introduction

Estimating the effect of a policy or treatment is a key question in econometrics and biostatistics. Randomization is the gold standard to avoid confounding, but cannot always be attained due to ethical or practical concerns. In the observational setting, we need to account for treatment assignment and adjust for confounding to obtain valid inference [63]. Besides the average treatment effect (ATE), we might also be interested in the conditional average treatment effect (CATE). The CATE indicates the expected benefit that an individual might have from a given treatment, given their observed covariates. It is especially of interest for making cost-effective policy choices and avoiding harm when treating a population. A medicine that might benefit the population on average might severely harm a certain subgroup [41].

Treatment heterogeneity is notoriously hard to estimate, as many randomized controlled trials (RCTs) are not powered to detect it [36, 41]. Furthermore, treatment heterogeneity can be sparse or nonexistent, calling for methods that can deal with this setting. Causal machine learning methods like R-, T-, and S-learners can struggle with smaller sample sizes, depending on the structure of the treatment effect and the underlying learner, as learning relationships becomes harder in data sets where signal is hard to find [53, 45]. Bayesian Causal Forests (BCF) have been shown to perform well in settings with weak treatment signals, as they are flexible tree ensemble methods whose priors on the tree structure and leaf values act as regularizers [34]. The choice of these priors is therefore central to controlling the estimated treatment heterogeneity.

BCF use separate prognostic and treatment-effect ensembles, which helps prevent the prognostic signal from distorting regularization of the treatment effect [33, 34]. The leaf prior in the treatment ensemble then determines how strongly regional effects are regularized. The BART prior assigns the same variance to every leaf, while the recently proposed Horseshoe Forests consider independent local scales [38]. These choices fix the relative variances to be equal or leave them unrelated, hence, they do not model how treatment-effect variation is distributed across the regions of a tree.

Joint variance allocation has proved useful in high-dimensional regression, where a total prior variance is distributed across coefficients to shrink negligible signals while retaining substantial ones [13, 77, 2, 76]. We adapt this principle to treatment trees through the Dirichlet-Tree Bayesian Causal Forest, DT-BCF. Our contributions are as follows:

  • We construct a topology-indexed family of priors that separates the total leaf variance of each treatment tree from its allocation across regions. The family contains equal division as a limiting allocation and Dirichlet-tree priors with random unequal shares.

  • We characterize the marginal shrinkage of the leaf values and derive a theoretical limit on how many regions can express substantial treatment-effect heterogeneity. We also establish posterior consistency.

  • We benchmark DT-BCF against Bayesian tree priors and external learners in controlled simulations, a causal-inference benchmark, and a real-data application. DT-BCF performs comparably to the Bayesian alternatives across a broad range of settings, with the largest gains when heterogeneity is concentrated in a few regions.

Related work

Bayesian tree models have been used for causal inference and treatment effect estimation [36, 31]. BCF separates the prognostic and treatment-effect functions so that each can receive its own regularization and includes the estimated propensity score in the prognostic function to mitigate regularization-induced confounding [33, 34]. Later extensions modify different parts of this construction. Targeted smooth BCF imposes smoothness along a designated effect modifier [66]. Shrinkage BCF places sparsity-inducing priors on the splitting probabilities to identify covariates that drive prognosis and treatment heterogeneity [17, 47]. Horseshoe Forests consider using the Horseshoe (HS) prior on the leaf nodes to allow different amounts of shrinkage across the covariate space [38]. DT-BCF changes the dependence among the leaf variances within each treatment tree, jointly regularizing regional contributions.

The closest prior constructions arise in joint shrinkage and variance decomposition. The Dirichlet–Laplace prior distributes a global scale across regression coefficients, while R2R^2 based priors allocate explained variance across model terms [13, 77, 2, 4, 76, 43]. Hierarchical decomposition priors allocate variance along a tree of additive model components specified by the analyst (Fuglstad et al., 2020). These approaches separate the overall amount of variation from its relative allocation among a fixed collection of model components. In DT-BCF, we allocate the variance by using the discovered topology of the tree recursively.

Outside Bayesian tree ensembles, the R-, X-learners, causal forests, multi-task Gaussian processes, and neural networks provide alternative estimators of heterogeneous effects [53, 45, 40, 11, 8, 65]. Our experiments emphasize comparisons among Bayesian tree priors and use the X- and R-learners as external references.

Methods

We first review the BCF treatment ensemble and identify the role of the leaf prior. We then show how to specify a tree prior which separates the total leaf variance of tree from its allocation across regions. To construct the allocation, we make use of the topology of tree. We also study the prior’s properties. Proofs are shown in Appendix § 11.

Bayesian Causal Forests

For unit ii, let Ai∈{0,1}A_i \in\{0,1\} denote treatment, xi∈Rp\boldsymbol{x}_i \in\mathbb{R}^p the observed covariates, and yi(a)y_i(a) the potential outcome under treatment a∈{0,1}a \in\{0,1\}. The observed response is yi=yi(Ai)y_i = y_i(A_i). We assume consistency and no interference, conditional exchangeability {yi(0),yi(1)}⊥ ⁣ ⁣ ⁣⊥Ai∣xi\{y_i(0), y_i(1)\} \perp\!\!\!\perp A_i \mid\boldsymbol{x}_i, and positivity 0<e(x)<10 < e(\boldsymbol{x}) < 1 on the covariate support, where e(x)=Pr⁡(Ai=1∣xi=x)e(\boldsymbol{x}) = \Pr(A_i = 1 \mid\boldsymbol{x}_i = \boldsymbol{x}) [37]. Under these conditions, the Conditional Average Treatment Effect (CATE)

τ(x)=E{yi(1)−yi(0)∣xi=x}=E(yi∣xi=x,Ai=1)−E(yi∣xi=x,Ai=0)(1)\tau(\boldsymbol{x}) = \mathbb{E}\{y_i(1)-y_i(0) \mid\boldsymbol{x}_i = \boldsymbol{x}\} = \mathbb{E}(y_i \mid\boldsymbol{x}_i = \boldsymbol{x}, A_i = 1) - \mathbb{E}(y_i \mid\boldsymbol{x}_i = \boldsymbol{x}, A_i = 0) \tag*{(1)}

is identified by the difference between the treated and control outcome regressions. Bayesian Causal Forests (BCF) model the observed outcome as

yi=μ(xi,e^i)+Ai{α+gτ(xi)}+εi,εi∼N(0,σ2),(2)y_i = \mu(\boldsymbol{x}_i,\widehat{e}_i) + A_i\{\alpha+ g_\tau(\boldsymbol{x}_i)\} + \varepsilon_i,\qquad\varepsilon_i \sim N(0,\sigma^2), \tag*{(2)}

where e^i\widehat{e}_i estimates e(xi)e(\boldsymbol{x}_i), μ\mu is the prognostic function, and τ(x)=α+gτ(x)\tau(\boldsymbol{x}) = \alpha+ g_\tau(\boldsymbol{x}) is the treatment-effect function [34]. Including the estimated propensity score in μ\mu helps protect the treatment-effect estimate from regularization-induced confounding [33]. We retain the BCF prior for μ\mu and modify only the leaf prior in the treatment ensemble gτg_\tau.

To define leaf priors we make use of the representation of a tree by its leaves. Let treatment tree TjT_j partition the covariate space into LjL_j leaves L(Tj)\mathcal{L}(T_j). Let zj(x)∈{0,1}Lj\boldsymbol{z}_j(\boldsymbol{x}) \in\{0,1\}^{L_j} be the one-hot vector indicating the leaf that contains x\boldsymbol{x}, and let hj=(hjℓ)ℓ∈L(Tj)\boldsymbol{h}_j = (h_{j\ell})_{\ell\in\mathcal{L}(T_j)} collect the leaf values. Tree jj contributes

g(x;Tj,hj)=zj(x)Thj,(3)g(\boldsymbol{x};T_j,\boldsymbol{h}_j) = \boldsymbol{z}_j(\boldsymbol{x})^\mathsf{T}\boldsymbol{h}_j, \tag*{(3)}

and gτ(x)=∑j=1mτg(x;Tj,hj)g_\tau(\boldsymbol{x}) = \sum_{j=1}^{m_\tau} g(\boldsymbol{x};T_j,\boldsymbol{h}_j). We consider Chipman et al. (2010)’s prior on the topology of a tree, in which a node at depth dd splits with probability ν(1+d)−η\nu(1+d)^{-\eta}. We select splitting variables using a discrete uniform distributions. Given TjT_j, BART assigns independent N(0,σh2)N(0,\sigma_h^2) distributions to the elements of hjh_j [21]. This leaves the choice of scale prior to determine how variation is distributed across the leaves. We consider common and leaf-specific scale priors next.

Problem Statement

Under BART, given that xx belongs to exactly one leaf, the prior variance of g(x;Tj,hj)g(x; T_j, h_j) is σh2\sigma_h^2. While considering further splits leaves the variance unchanged, the total variance the leaves represent increases. For a treatment tree, we seek two properties. 1) Leaf variances should be able to differ across covariate space, so that a leaf is shrunk strongly toward zero where the data supports no heterogeneity. 2) Second, the sum of the leaf variances of a tree should have a prior that does not depend on LjL_j, so that a split redistributes prior variance within a tree (i.e the total variability of a tree) instead of adding to it.

Neither property holds under the BART leaf prior. Priors on the leafs with conditionally independent local scales, such as the Horseshoe [19, 57, 54], meet the first but not the second requirement. Under any such prior, the sum of the leaf variances is increasing in LjL_j, implying that the second property cannot be satisfied. Scaling BART leaf variances by LjL_j could endow the prior with the second property, but every leaf would receive the same fixed variance, offering no adaptability. We meet both by letting the law of the leaf variances depend on the topology of the tree.

To this end, we propose a prior that separates the total variance of a tree from its allocation across leaves. Tree jj receives a positive tree scale ψj2\psi_j^2 and an allocation vector ϕj=(ϕjℓ)ℓ∈L(Tj)\phi_j=(\phi_{j\ell})_{\ell\in\mathcal{L}(T_j)} satisfying ϕjℓ>0\phi_{j\ell}>0 and ∑ℓϕjℓ=1\sum_{\ell}\phi_{j\ell}=1. The prior on ψj2\psi_j^2 is independent of the tree, while the law of ϕj\phi_j is constructed using information from the topology. Given TjT_j, ψj2\psi_j^2 and ϕj\phi_j, the leaf values are independent with hjℓ∼N(0,Vjℓ)h_{j\ell}\sim\mathrm{N}(0,V_{j\ell}), with Vjℓ=VjϕjℓV_{j\ell}=V_j\phi_{j\ell} and Vj=σ2ωψj2V_j=\sigma^2\omega\psi_j^2, where ω>0\omega>0 is a shared constant of the forest. Since, ∑ℓ∈L(Tj)Vjℓ=Vj\sum_{\ell\in\mathcal{L}(T_j)}V_{j\ell}=V_j for every tree jj and every value of LjL_j, we immediately satisfy the second property we seek. We are therefore able to jointly control the total amount of signal that the tree can represent. The compositional nature of VjℓV_{j\ell} implies that a larger share for one leaf leaves less total variance VjV_j for the others. This is precisely the joint regularization of leaf values we aim to achieve. To achieve this we need to specify prior distributions for ψj2\psi_j^2 and ϕj\phi_j.

Dirichlet-Tree Allocation

We now specify the law of the shares ϕj\phi_j given the topology of the tree. The random topology of TjT_j requires a law for the shares on every possible tree. A law chosen separately for each topology would not be able to express how the shares of a tree relate to those of a tree obtained by splitting one of its leaves. Therefore, we define the shares ϕj\phi_j through the same binary recursion that constructs the partition. The main idea is that each internal node divides the share of total variance VjV_j it receives between its two children, while the rest of the tree remains unchanged.

Let I(Tj)\mathcal{I}(T_j) be the internal nodes of TjT_j, and ϕjv\phi_{jv} the share reaching node vv, with ϕjv=1\phi_{jv}=1 at the root. Let the branch fraction Ujv∈(0,1)U_{jv}\in(0,1) at node v∈I(Tj)v\in\mathcal{I}(T_j) send

ϕjvL=Ujvϕjv,ϕjvR=(1−Ujv)ϕjv(4)\phi_{jv_L}=U_{jv}\phi_{jv},\qquad\phi_{jv_R}=(1-U_{jv})\phi_{jv} \tag*{(4)}

to the left and right children vLv_L and vRv_R. Therefore, each leaf share is the product of the branch fractions along its root-to-leaf path, using UjvU_{jv} on a left branch and 1−Ujv1-U_{jv} on a right branch.

If the branch fractions UjvU_{jv} are independent with

References

  1. [1]Agarwal, S., Dutta, S., and Mukhopadhyay, M. (2025). Bayesian Variable Selection Under High-Dimensional Settings with Grouped Covariates. Bayesian Analysis. https://doi.org/10.1214/25-BA1525 (14).
  2. [2]Aguilar, J. E. and Bürkner, P.-C. (2023). Intuitive Joint Priors for Bayesian Linear Multilevel Models: The R2D2M2 Prior. Electronic Journal of Statistics 17(1), 1711–1767. https://doi.org/10.1214/23-EJS2136 (2, 3, 8, 14, 29, 30).
  3. [3]Aguilar, J. E. and Bürkner, P.-C. (2025a). Dependency-Aware Shrinkage Priors for High Dimensional Regression. https://arxiv.org/abs/2505.10715 (7).
  4. [4]Aguilar, J. E. and Bürkner, P.-C. (2025b). Generalized Decomposition Priors on R2. Bayesian Analysis, 1–34. https://doi.org/10.1214/25-BA1524 (3, 14).
  5. [5]Aguilar, J. E., Kohns, D., Vehtari, A., and Bürkner, P.-C. (2025). R2 priors for Grouped Variance Decomposition in High-dimensional Regression. arXiv preprint arXiv:2507.11833 (14).arxiv.org/abs/2507.11833
  6. [6]Aitchison, J. and Shen, S. M. (1980). Logistic-Normal Distributions: Some Properties and Uses. Biometrika 67(2), 261–272. http://www.jstor.org/stable/2335470 (14).DOI
  7. [7]Aitchison, J. (1986). The statistical analysis of compositional data. Monographs on statistics and applied probability (Series). Chapman and Hall (5).
  8. [8]Alaa, A. M. and van der Schaar, M. (2017). Bayesian Inference of Individualized Treatment Effects Using Multi-Task Gaussian Processes. In: Advances in Neural Information Processing Systems. Vol. 30 (3).arxiv.org/abs/1704.02801
  9. [9]Andrews, D. F. and Mallows, C. L. (1974). Scale Mixtures of Normal Distributions. Journal of the Royal Statistical Society: Series B (Methodological) 36(1), 99–102. https://doi.org/10.1111/j.2517-6161.1974.tb00989.x (34).
  10. [10]Angrist, J. D., Imbens, G. W., and Rubin, D. B. (1996). Identification of Causal Effects Using Instrumental Variables. Journal of the American Statistical Association 91(434), 444–455. https://doi.org/10.1080/01621459.1996.10476902 (14).
  11. [11]Athey, S., Tibshirani, J., and Wager, S. (2019). Generalized random forests. The Annals of Statistics 47(2), 1148–1178 (3).arxiv.org/abs/1610.01271
  12. [12]Bai, R. and Ghosh, M. (2019). Large-scale multiple hypothesis testing with the normal-beta prime prior. Statistics 53(6), 1210–1233. https://doi.org/10.1080/02331888.2019.1662017 (7).
  13. [13]Bhattacharya, A., Pati, D., Pillai, N. S., and Dunson, D. B. (2015). Dirichlet–Laplace Priors for Optimal Shrinkage. Journal of the American Statistical Association 110(512), 1479–1490. https://doi.org/10.1080/01621459.2014.960967 (2, 3).
  14. [14]Boss, J., Datta, J., Wang, X., Park, S. K., Kang, J., and Mukherjee, B. (2023). Group Inverse-Gamma Gamma Shrinkage for Sparse Linear Models with Block-Correlated Regressors. Bayesian Analysis, 1–30. https://doi.org/10.1214/23-BA1371 (14).
  15. [15]Cancer Genome Atlas Research Network (2017). Integrated Genomic Characterization of Pancreatic Ductal Adenocarcinoma. Cancer Cell 32(2), 185–203.e13. https://doi.org/10.1016/j.ccell.2017.07.007 (12, 41, 62).
  16. [16]Caron, A., Baio, G., and Manolopoulou, I. (2022a). Estimating individual treatment effects using non-parametric regression models. Journal of the Royal Statistical Society Series A: Statistics in Society 185(3), 1115–1149. https://doi.org/10.1111/rssa.12824 (25).
  17. [17]Caron, A., Baio, G., and Manolopoulou, I. (2022b). Shrinkage Bayesian Causal Forests for Heterogeneous Treatment Effects Estimation. Journal of Computational and Graphical Statistics 31(4), 1202–1214. https://doi.org/10.1080/10618600.2022.2067549 (2, 14, 24).
  18. [18]Carvalho, C. M., Polson, N. G., and Scott, J. G. (2009). Handling Sparsity via the Horseshoe. In: Proceedings of the Twelfth International Conference on Artificial Intelligence and Statistics. Ed. by D. van Dyk and M. Welling. Vol. 5. Proceedings of Machine Learning Research. PMLR, 73–80. https://proceedings.mlr.press/v5/carvalho09a.html (9).
  19. [19]Carvalho, C. M., Polson, N. G., and Scott, J. G. (2010). The Horseshoe Estimator for Sparse Signals. Biometrika 97(2), 465–480. https://doi.org/10.1093/biomet/asq017 (4, 35).
  20. [20]Chen, T. and Guestrin, C. (2016). XGBoost: A Scalable Tree Boosting System. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. KDD ’16. San Francisco, California, USA: Association for Computing Machinery, 785–794. https://doi.org/10.1145/2939672.2939785 (9, 25, 40, 41).
  21. [21]Chipman, H. A., George, E. I., and McCulloch, R. E. (2010). BART: Bayesian Additive Regression Trees. The Annals of Applied Statistics 4(1), 266–298. https://doi.org/10.1214/09-AOAS285 (3, 4, 21, 22, 27, 28).
  22. [23]Dorie, V., Hill, J., Shalit, U., Scott, M., and Cervone, D. (2019). Automated versus Do-It-Yourself Methods for Causal Inference: Lessons Learned from a Data Analysis Competition. Statistical Science 34(1), 43–68. https://doi.org/10.1214/18-STS667 (12, 41, 59, 60).
  23. [24]Eddelbuettel, D. and François, R. (2011). Rcpp: Seamless R and C++ Integration. Journal of Statistical Software 40(8), 1–18. https://doi.org/10.18637/jss.v040.i08 (41).
  24. [25]Feller, W. (1971). An Introduction to Probability Theory and Its Applications, Volume II. Second. New York: John Wiley & Sons (7).
  25. [26]Friedman, J., Hastie, T., and Tibshirani, R. (2010). Regularization Paths for Generalized Linear Models via Coordinate Descent. Journal of Statistical Software 33(1), 1–22. https://doi.org/10.18637/jss.v033.i01 (39–41).
  26. [28]Fuglstad, G.-A., Simpson, D., Lindgren, F., and Rue, H. (2019). Constructing Priors that Penalize the Complexity of Gaussian Random Fields. Journal of the American Statistical Association 114(525), 445–452. https://doi.org/10.1080/01621459.2017.1415907 (14).
  27. [29]Gelman, A., Vehtari, A., Simpson, D., Margossian, C. C., Carpenter, B., Yao, Y., Kennedy, L., Gabry, J., Bür kner, P.-C., and Modrák, M. (2020). Bayesian Workflow. arXiv:2011.01808. https://doi.org/10.48550/ARXIV.2011.01808 (6).
  28. [30]Ghosal, S., Ghosh, J. K., and van der Vaart, A. W. (2000). Convergence rates of posterior distributions. The Annals of Statistics 28(2), 500–531. https://doi.org/10.1214/aos/1016218228 (36).
  29. [31]Green, D. P. and Kern, H. L. (2012). Modeling Heterogeneous Treatment Effects in Survey Experiments with Bayesian Additive Regression Trees. Public Opinion Quarterly 76(3), 491–511. https://doi.org/10.1093/poq/nfs036 (2).
  30. [32]Grinsztajn, L., Oyallon, E., and Var oquaux, G. (2022). Why do tree-based models still outperform deep learning on typical tabular data? In: Advances in Neural Information Processing Systems. Vol. 35, 507–520. https://proceedings.neurips.cc/paper_files/paper/2022/hash/0378c7692da36807bdcec87ab043cdadc-Abstract-Datasets_and_Benchmarks.html (25).
  31. [33]Hahn, P. R., Carvalho, C. M., Puelz, D., and He, J. (2018). Regularization and Confounding in Linear Regression for Treatment Effect Estimation. Bayesian Analysis 13(1), 163–182. https://doi.org/10.1214/16-BA1044 (2, 3, 14, 23).
  32. [34]Hahn, P. R., Murray, J. S., and Carvalho, C. M. (2020). Bayesian Regression Tree Models for Causal Inference: Regularization, Confounding, and Heterogeneous Effects. Bayesian Analysis 15(3), 965–1056. https://doi.org/10.1214/19-BA1195 (2, 3, 8, 9, 14, 22, 23, 26, 27, 39, 41).
  33. [35]Hastie, T. and Tibshirani, R. (2000). Bayesian backfitting (with comments and a rejoinder by the authors. Statistical Science 15(3), 196–223 (8, 21, 23).DOI
  34. [36]Hill, J. L. (2011). Bayesian Nonparametric Modeling for Causal Inference. Journal of Computational and Graphical Statistics 20(1), 217–240. https://doi.org/10.1198/jcgs.2010.08162 (2, 40).
  35. [37]Imbens, G. W. and Rubin, D. B. (2015). Causal Inference for Statistics, Social, and Biomedical Sciences: An Introduction. Cambridge: Cambridge University Press (3).
  36. [38]Jacobs, T., van Wieringen, W. N., and Pas, S. L. v. d. (2026). Horseshoe Forests for High-Dimensional Causal Structure Learning. Bayesian Analysis. Advance publication. https://doi.org/10.1214/26-BA1603 (2, 9, 12, 24, 26, 39, 41, 62).
  37. [39]Kalbfleisch, J. D. and Prentice, R. L. (2002). The Statistical Analysis of Failure Time Data. 2nd. Vol. 460. Hoboken, NJ: John Wiley & Sons (26).
  38. [40]Kennedy, E. H. (2023). Towards optimal doubly robust estimation of heterogeneous causal effects. Electronic Journal of Statistics 17(2), 3008–3049. https://doi.org/10.1214/23-EJS2157 (3, 25).
  39. [41]Kent, D. M., Steyerberg, E., and van Klaveren, D. (2018). Personalized Evidence Based Medicine: Predictive Approaches to Heterogeneous Treatment Effects. BMJ 363, k4245 (1, 2).DOI
  40. [42]Klippert, D., Friedrich, S., and Pauly, M. (2026). A Guide to Estimating Conditional Average Treatment Effects in Competing Risks Settings. arXiv:2606.18281. https://arxiv.org/abs/2606.18281 (25).
  41. [43]Kohns, D., Kallioinen, N., McLatchie, Y., and Vehtari, A. (2026). The ARR2 prior: flexible predictive prior definition for Bayesian auto-regressions. Bayesian Analysis 21(3), 1131–1162 (3).DOI
  42. [44]Kruijer, W., Rousseau, J., and van der Vaart, A. W. (2010). Adaptive Bayesian Density Estimation with Location-Scale Mixtures. Electronic Journal of Statistics 4. https://doi.org/10.1214/10-EJS584 (29, 32).
  43. [45]Künzel, S. R., Sekhon, J. S., Bickel, P. J., and Yu, B. (2019). Metalearners for estimating heterogeneous treatment effects using machine learning. Proceedings of the National Academy of Sciences 116(10), 4156–4165. https://doi.org/10.1073/pnas.1804597116 (2, 3, 9, 14, 25, 40).
  44. [46]Lin, J. (2016). On the Dirichlet distribution. Department of Mathematics and Statistics, Queens University 40, 46 (5, 32).
  45. [47]Linero, A. R. (2018). Bayesian Regression Trees for High-Dimensional Prediction and Variable Selection. Journal of the American Statistical Association 113(522), 626–636. https://doi.org/10.1080/01621459.2016.1264957 (2, 9, 24, 28, 39).
  46. [48]Linero, A. R. (2024). In Nonparametric and High-Dimensional Models, Bayesian Ignorability is an Informative Prior. Journal of the American Statistical Association 119(548), 2785–2798. https://doi.org/10.1080/01621459.2023.2278202 (23).
  47. [49]Manderson, A. A. and Goudie, R. J. B. (2026). Translating predictive distributions into informative priors. Journal of the American Statistical Association 0(ja), 1–18. https://doi.org/10.1080/01621459.2026.2614034 (6).
  48. [50]Minka, T. P. (2004). The Dirichlet-tree Distribution. Tech. rep. Originally circulated in 1999, revised October 2004. Justsystem Pittsburgh Research Center. https://tminka.github.io/papers/dirichlet/minka-dirtree.pdf (5).
  49. [51]National Cancer Institute (2014). The Cancer Genome Atlas Program: Human Subjects Protection and Data Access Policies. https://www.cancer.gov/about-nci/organization/ccg/research/structural-genomics/tcga/history/policies/tcga-human-subjects-data-policies.pdf. Revision of 16 January 2014 (62).
  50. [52]Neal, R. M. (2003). Slice Sampling. The Annals of Statistics 31(3), 705–767 (8, 29).
  51. [53]Nie, X. and Wager, S. (2021). Quasi-oracle estimation of heterogeneous treatment effects. Biometrika 108(2), 299–319 (2, 3, 9, 14, 25, 26, 40, 41).arxiv.org/abs/1712.04912
  52. [54]Pas, S. L. v. d. (2021). Theoretical guarantees for the horseshoe and other global-local shrinkage priors. In: Handbook of Bayesian Variable Selection. Chapman and Hall/CRC, 133–160 (4).DOI
  53. [55]Pas, S. L. v. d., Kleijn, B. J. K., and Vaart, A. W. v. d. (2014). The horseshoe estimator: Posterior concentration around nearly black vectors. Electronic Journal of Statistics 8(2), 2585–2618. https://doi.org/10.1214/14-EJS962 (6).
  54. [56]Pas, S. L. v. d., Szabó, B., and Vaart, A. W. v. d. (2017). Adaptive posterior contraction rates for the horseshoe. Electronic Journal of Statistics 11(2). https://doi.org/10.1214/17-EJS1316 (6).
  55. [57]Piironen, J. and Vehtari, A. (2017). Sparsity Information and Regularization in the Horseshoe and Other Shrinkage Priors. Electronic Journal of Statistics 11(2), 5018–5051. https://doi.org/10.1214/17-EJS1337SI (4, 7).
  56. [58]Pratola, M. T. (2016). Efficient Metropolis–Hastings Proposal Mechanisms for Bayesian Regression Tree Models. Bayesian Analysis 11(3), 885–911. https://doi.org/10.1214/16-BA999 (8, 22).
  57. [59]R Core Team (2026). R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. Vienna, Austria. https://www.R-project.org/ (41).
  58. [60]Robert, C. P. (1995). Simulation of Truncated Normal Variables. Statistics and Computing 5(2), 121–125 (31).
  59. [61]Robinson, P. M. (1988). Root-N-Consistent Semiparametric Regression. Econometrica 56(4), 931–954. http://www.jstor.org/stable/1912705 (14).DOI
  60. [62]Ročková, V. and Saha, E. (2019). On Theory for BART. In: Proceedings of the Twenty-Second International Conference on Artificial Intelligence and Statistics. Vol. 89. Proceedings of Machine Learning Research, 2839–2848. https://proceedings.mlr.press/v89/rockova19a.html (7).
  61. [63]Rosenbaum, P. R. and Rubin, D. B. (1983). The Central Role of the Propensity Score in Observational Studies for Causal Effects. Biometrika 70(1), 41–55 (1).
  62. [64]Schwartz, L. (1965). On Bayes Procedures. Zeitschrift für Wahrscheinlichkeitstheorie und Verwandte Gebiete 4(1), 10–26. https://doi.org/10.1007/BF00553796 (36, 38).
  63. [65]Shalit, U., Johansson, F. D., and Sontag, D. (2017). Estimating Individual Treatment Effect: Generalization Bounds and Algorithms. In: Proceedings of the 34th International Conference on Machine Learning. Vol. 70. Proceedings of Machine Learning Research, 3076–3085 (3).arxiv.org/abs/1606.03976
  64. [66]Starling, J. E., Murray, J. S., Lohr, P. A., Aiken, A. R. A., Carvalho, C. M., and Scott, J. G. (2021). Targeted Smooth Bayesian Causal Forests: An Analysis of Heterogeneous Treatment Effects for Simultaneous versus Interval Medical Abortion Regimens over Gestation. The Annals of Applied Statistics 15(3), 1194–1219. https://doi.org/10.1214/20-AOAS1438 (2, 14, 24).
  65. [67]Tew, S. Y., Schmidt, D. F., and Boley, M. (2025). GRASP: Grouped Regression with Adaptive Shrinkage Priors. https://arxiv.org/abs/2506.18092 (14).
  66. [68]Therneau, T. M. (2024). A Package for Survival Analysis in R. R package version 3.8-12. https://CRAN.R-project.org/package=survival (39, 41).
  67. [69]Torre-Healy, L. A., Kawalerski, R. R., Oh, K., Chrastecka, L., Peng, X. L., Aguirre, A. J., Rashid, N. U., Yeh, J. J., and Moffitt, R. A. (2023). Open-Source Curation of a Pancreatic Ductal Adenocarcinoma Gene Expression Analysis Platform (pdacR) Supports a Two-Subtype Model. Communications Biology 6, 163. https://doi.org/10.1038/s42003-023-04461-6 (41, 62).
  68. [71]Vehtari, A., Gelman, A., and Gabry, J. (2017). Practical Bayesian model evaluation using leave-one-out cross-validation and WAIC. Statistics and Computing 27(5), 1413–1432 (40, 63).DOI
  69. [72]Vehtari, A., Gelman, A., Simpson, D., Carpenter, B., and Bürкner, P.-C. (2021). Rank-Normalization, Folding, and Localization: An Improved \widehat{R} for Assessing Convergence of MCMC (with Discussion). Bayesian Analysis 16(2), 667–718. https://doi.org/10.1214/20-BA1221 (9).
  70. [73]Vehtari, A., Simpson, D., Gelman, A., Yao, Y., and Gabry, J. (2024). Pareto Smoothed Importance Sampling. Journal of Machine Learning Research 25(72), 1–58 (63).
  71. [74]West, M. (1987). On scale mixtures of normal distributions. Biometrika 74(3), 646–648. https://doi.org/10.1093/biomet/74.3.646 (34).
  72. [76]Yanchenko, E., Bondell, H. D., and Reich, B. J. (2024). The R2D2 Prior for Generalized Linear Mixed Models. https://arxiv.org/abs/2111.10718 (2, 3).
  73. [77]Zhang, Y. D., Naughton, B. P., Bondell, H. D., and Reich, B. J. (2022). Bayesian Regression Using a Prior on the Model Fit: The R2-D2 Shrinkage Prior. Journal of the American Statistical Association 117(538), 862–874. https://doi.org/10.1080/01621459.2020.1825449 (2, 3, 8, 14, 29, 30).

Paper details

vlnc:W_e4zm7euw
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents