Technology Category
Machine Learning: Bayesian LearningKnowledge Representation and Reasoning: Description LogicsData Mining & Knowledge Management: Scalability, Parallel & Distributed Systems
Structured Bayesian Regression Tree Models for Estimating Distributed Lag Effects: The R Package dlmtree
🤖 AI Summary
In environmental and epidemiological studies, exposure–outcome associations often exhibit time lags, with lagged effects characterized by smoothness, nonlinearity, and subgroup heterogeneity. Conventional distributed lag models (DLMs) lack flexibility to capture such complex patterns. To address this, we propose dlmtree—a novel Bayesian regression tree–driven DLM framework. dlmtree innovatively incorporates tree-based structures into DLMs, enabling adaptive, piecewise-smooth estimation of lagged effects while simultaneously identifying heterogeneous subgroups. Implemented as a CRAN-certified R package, dlmtree provides integrated functionality for model specification, Bayesian inference, visualization, and heterogeneity analysis. The framework substantially enhances interpretability, flexibility, and practical utility in estimating time-varying causal effects from longitudinal or time-series data.
Technology Category
Application Category
📝 Abstract
When examining the relationship between an exposure and an outcome, there is often a time lag between exposure and the observed effect on the outcome. A common statistical approach for estimating the relationship between the outcome and lagged measurements of exposure is a distributed lag model (DLM). Because repeated measurements are often autocorrelated, the lagged effects are typically constrained to vary smoothly over time. A recent statistical development on the smoothing constraint is a tree structured DLM framework. We present an R package dlmtree, available on CRAN, that integrates tree structured DLM and extensions into a comprehensive software package with user-friendly implementation. A conceptual background on tree structured DLMs and demonstration of the fitting process of each model using simulated data are provided. We also demonstrate inference and interpretation using the fitted models, including summary and visualization. Additionally, a built-in shiny app for heterogeneity analysis is included.
Problem
Research questions and friction points this paper is trying to address.
Estimating distributed lag effects in exposure-outcome relationships
Addressing autocorrelation in lagged exposure measurements
Providing user-friendly software for tree-structured DLMs
Innovation
Methods, ideas, or system contributions that make the work stand out.
Tree structured DLM framework for lag effects
R package dlmtree integrates user-friendly implementation
Includes shiny app for heterogeneity analysis
🔎 Similar Papers
2024-08-29arXiv.orgCitations: 0
