Predicting abnormal capital flow episodes with machine learning methods

doi:10.1016/j.qref.2025.102026

Predicting abnormal capital flow episodes with machine learning methods

2025-06-24

0

PRE

AI

DOI：10.1016/j.qref.2025.102026

Original

Original request for help

Save

Abstract

En 中文

• Introduces early warning framework for abnormal capital flows using machine learning. • Evaluates linear and ML models, showing tree ensembles’ superior predictive accuracy. • Uses Shapley decomposition to identify and explain key predictors, like DLD. • Offers guidance for policymakers to formulate policy against volatile capital flows.

Predicting abnormal capital flow episodes with machine learning methods

Abstract

Journal

Organization