This dissertation centers on Bayesian Additive Regression Trees (BART) and Accelerated BART (XBART) and presents a series of models that tackle extrapolation, classification, and causal inference challenges. To improve extrapolation in tree-based models, I propose a method called local Gaussian Process (GP) that combines Gaussian process regression with trained BART trees. This allows for extrapolation based on the most relevant data points and covariate variables determined by the trees' structure. The local GP technique is extended to the Bayesian causal forest (BCF) models to address the positivity violation issue in causal inference. Additionally, I introduce the LongBet model to estimate time-varying, heterogeneous treatment effects in panel data. Furthermore, I present a Poisson-based model, with a modified likelihood for XBART for the multi-class classification problem.

This dissertation centers on treatment effect estimation in the field of causal inference, and aims to expand the toolkit for effect estimation when the treatment variable is binary. Two new stochastic tree-ensemble methods for treatment effect estimation in the continuous outcome setting are presented. The Accelerated Bayesian Causal Forrest (XBCF) model handles variance via a group-specific parameter, and the Heteroskedastic version of XBCF (H-XBCF) uses a separate tree ensemble to learn covariate-dependent variance. This work also contributes to the field of survival analysis by proposing a new framework for estimating survival probabilities via density regression. Within this framework, the Heteroskedastic Accelerated Bayesian Additive Regression Trees (H-XBART) model, which is also developed as part of this work, is utilized in treatment effect estimation for right-censored survival outcomes. All models have been implemented as part of the XBART R package, and their performance is evaluated via extensive simulation studies with appropriate sets of comparators. The contributed methods achieve similar levels of performance, while being orders of magnitude (sometimes as much as 100x) faster than comparator state-of-the-art methods, thus offering an exciting opportunity for treatment effect estimation in the large data setting.

This dissertation covers several topics in machine learning and causal inference. First, the question of “feature selection,” a common byproduct of regularized machine learning methods, is investigated theoretically in the context of treatment effect estimation. This involves a detailed review and extension of frameworks for estimating causal effects and in-depth theoretical study. Next, various computational approaches to estimating causal effects with machine learning methods are compared with these theoretical desiderata in mind. Several improvements to current methods for causal machine learning are identified and compelling angles for further study are pinpointed. Finally, a common method used for “explaining” predictions of machine learning algorithms, SHAP, is evaluated critically through a statistical lens.

This dissertation develops versatile modeling tools to estimate causal effects when conditional unconfoundedness is not immediately satisfied. Chapter 2 provides a brief overview ofcommon techniques in causal inference, with a focus on models relevant to the data explored in later chapters. The rest of the dissertation focuses on the development of novel “reduced form” models which are designed to assess the particular challenges of different datasets. Chapter 3 explores the question of whether or not forecasts of bankruptcy cause bankruptcy. The question arises from the observation that companies issued going concern opinions were more likely to go bankrupt in the following year, leading people to speculate that the opinions themselves caused the bankruptcy via a “self-fulfilling prophecy”. A Bayesian machine learning sensitivity analysis is developed to answer this question. In exchange for additional flexibility and fewer assumptions, this approach loses point identification of causal effects and thus a sensitivity analysis is developed to study a wide range of plausible scenarios of the causal effect of going concern opinions on bankruptcy. Reported in the simulations are different performance metrics of the model in comparison with other popular methods and a robust analysis of the sensitivity of the model to mis-specification. Results on empirical data indicate that forecasts of bankruptcies likely do have a small causal effect. Chapter 4 studies the effects of vaccination on COVID-19 mortality at the state level in the United States. The dynamic nature of the pandemic complicates more straightforward regression adjustments and invalidates many alternative models. The chapter comments on the limitations of mechanistic approaches as well as traditional statistical methods to epidemiological data. Instead, a state space model is developed that allows the study of the ever-changing dynamics of the pandemic’s progression. In the first stage, the model decomposes the observed mortality data into component surges, and later uses this information in a semi-parametric regression model for causal analysis. Results are investigated thoroughly for empirical justification and stress-tested in simulated settings.

Tracking disease cases is an essential task in public health; however, tracking the number of cases of a disease may be difficult not every infection can be recorded by public health authorities. Notably, this may happen with whole country measles case reports, even such countries with robust registration systems. Eilertson et al. (2019) propose using a state-space model combined with maximum likelihood methods for estimating measles transmission. A Bayesian approach that uses particle Markov Chain Monte Carlo (pMCMC) is proposed to estimate the parameters of the non-linear state-space model developed in Eilertson et al. (2019) and similar previous studies. This dissertation illustrates the performance of this approach by calculating posterior estimates of the model parameters and predictions of the unobserved states in simulations and case studies. Also, Iteration Filtering (IF2) is used as a support method to verify the Bayesian estimation and to inform the selection of prior distributions. In the second half of the thesis, a birth-death process is proposed to model the unobserved population size of a disease vector. This model studies the effect of a disease vector population size on a second affected population. The second population follows a non-homogenous Poisson process when conditioned on the vector process with a transition rate given by a scaled version of the vector population. The observation model also measures a potential threshold event when the host species population size surpasses a certain level yielding a higher transmission rate. A maximum likelihood procedure is developed for this model, which combines particle filtering with the Minorize-Maximization (MM) algorithm and extends the work of Crawford et al. (2014).

Spatial regression is one of the central topics in spatial statistics. Based on the goals, interpretation or prediction, spatial regression models can be classified into two categories, linear mixed regression models and nonlinear regression models. This dissertation explored these models and their real world applications. New methods and models were proposed to overcome the challenges in practice. There are three major parts in the dissertation.

In the first part, nonlinear regression models were embedded into a multistage workflow to predict the spatial abundance of reef fish species in the Gulf of Mexico. There were two challenges, zero-inflated data and out of sample prediction. The methods and models in the workflow could effectively handle the zero-inflated sampling data without strong assumptions. Three strategies were proposed to solve the out of sample prediction problem. The results and discussions showed that the nonlinear prediction had the advantages of high accuracy, low bias and well-performed in multi-resolution.

In the second part, a two-stage spatial regression model was proposed for analyzing soil carbon stock (SOC) data. In the first stage, there was a spatial linear mixed model that captured the linear and stationary effects. In the second stage, a generalized additive model was used to explain the nonlinear and nonstationary effects. The results illustrated that the two-stage model had good interpretability in understanding the effect of covariates, meanwhile, it kept high prediction accuracy which is competitive to the popular machine learning models, like, random forest, xgboost and support vector machine.

A new nonlinear regression model, Gaussian process BART (Bayesian additive regression tree), was proposed in the third part. Combining advantages in both BART and Gaussian process, the model could capture the nonlinear effects of both observed and latent covariates. To develop the model, first, the traditional BART was generalized to accommodate correlated errors. Then, the failure of likelihood based Markov chain Monte Carlo (MCMC) in parameter estimating was discussed. Based on the idea of analysis of variation, back comparing and tuning range, were proposed to tackle this failure. Finally, effectiveness of the new model was examined by experiments on both simulation and real data.