FIELD NOTES / Clinical Data Science

Clinical Machine Learning & Causal Inference

PublishedEvidence A

How can heterogeneous clinical observations become interpretable predictions and explicitly qualified causal estimates?

Last reviewed 2026-10-19 · Public research / code

Overview

Mental health and clinical outcomes · Clinical Machine Learning & Causal Inference

从精神健康临床问题出发，连接可解释预测、因果假设、网络与心理测量。

Research Question

How can heterogeneous clinical observations become interpretable predictions and explicitly qualified causal estimates?

Technical Approach

Published work spans DAG-guided observational causal modeling, interpretable machine learning, family network analysis and psychometric validation. A separate public AIPW pipeline checks DAG validity and reports aggregate diagnostics.

PythonRDoWhy (published study)DAGsscikit-learnSHAPlavaanNetwork analysis

My Contribution

First-listed author of the school-aversion causal study and family network study; second-listed author of the NSSI paper and co-first author of the quantitative imaging paper (author-confirmed). Public DAG-Dowhy code provides an inspectable engineering example; detailed CRediT roles are left unspecified.

Evidence

Peer-reviewed papers with verified DOI metadata and author order.
Local school-aversion reproduction entry and analysis modules inspected.
Public DAG-Dowhy pipeline, synthetic tests and CI.

Public code ↗Published evidence ↗

Outputs

Peer-reviewed causal modeling, ML and network papers
Public aggregate-only causal analysis pipeline

Current Status