Work

Machine learning-driven approach to designing carbon-based electrocatalysts for oxygen reduction reaction

Year: 2026

Type: article

Abstract: A machine learning-driven approach was used to evaluate a carbon-based electrocatalyst for H2O2 synthesis via a twoelectron oxygen reduction reaction. The R2 value of the optimal preliminary model, using the eXtreme Gradient Boosting (XGB) algorithm with only four H2O2 selectivity descriptors, was 0.61. The dilemma of small data for machine learning was demonstrated, and a simple strategy for overcoming it was proposed to provide a new perspective on catalyst design.

Source: Mendeleev Communications

DOI: 10.71267/mencom.8059