Abstract:In many fields of scientific research and real-world applications, unbiased estimation of causal effects from non-experimental data is crucial for understanding the mechanism underlying the data and for decision-making on effective responses or interventions. A great deal of research has been conducted to address this challenging problem from different angles. For estimating causal effect in observational data, assumptions such as Markov condition, faithfulness, and causal sufficiency are always made. Under the assumptions, full knowledge, such as a set of covariates or an underlying causal graph, is typically required. A practical challenge is that in many applications, no such full knowledge or only some partial knowledge is available. In recent years, research has emerged to use search strategies based on graphical causal modelling to discover useful knowledge from data for causal effect estimation, with some mild assumptions, and has shown promise in tackling the practical challenge. In this survey, we review these data-driven methods on causal effect estimation for a single treatment with a single outcome of interest and focus on the challenges faced by data-driven causal effect estimation. We concisely summarise the basic concepts and theories that are essential for data-driven causal effect estimation using graphical causal modelling but are scattered around the literature. We identify and discuss the challenges faced by data-driven causal effect estimation and characterise the existing methods by their assumptions and the approaches to tackling the challenges. We analyse the strengths and limitations of the different types of methods and present an empirical evaluation to support the discussions. We hope this review will motivate more researchers to design better data-driven methods based on graphical causal modelling for the challenging problem of causal effect estimation.

Mixed Graphical Models for Causal Analysis of Multi-modal Variables

Learning Cluster Causal Diagrams: an Information-Theoretic Approach

Towards Human-like Perception: Learning Structural Causal Model in Heterogeneous Graph

High-Dimensional Undirected Graphical Models for Arbitrary Mixed Data

Directed Cyclic Graph for Causal Discovery from Multivariate Functional Data

A cautious approach to constraint-based causal model selection

Continuous Bayesian Model Selection for Multivariate Causal Discovery

Visual Analysis of Multi-outcome Causal Graphs

Identification of Latent Variables From Graphical Model Residuals

High-dimensional Mixed Graphical Models

Causal disentanglement of multimodal data

Review of Causal Discovery Methods Based on Graphical Models

A Hybrid Causal Structure Learning Algorithm for Mixed-Type Data

Causal Discovery over High-Dimensional Structured Hypothesis Spaces with Causal Graph Partitioning

Causal Graph Discovery from Self and Mutually Exciting Time Series

Causal Inference in Directed, Possibly Cyclic, Graphical Models

Scalable and Flexible Causal Discovery with an Efficient Test for Adjacency

Kernel-Based Differentiable Learning of Non-Parametric Directed Acyclic Graphical Models

Characterization of causal ancestral graphs for time series with latent confounders

On Learning Necessary and Sufficient Causal Graphs

Data-Driven Causal Effect Estimation Based on Graphical Causal Modelling: A Survey