Work

Code for: An Asymmetric Permeability Signature Underlies CO₂/CH₄ Selectivity in Polymeric Membranes: An Interpretable Machine Learning Analysis

Year: 2026

Type: software

Abstract: This repository contains the complete computational pipeline used to develop and interpret a Random Forest model for predicting CO₂/CH₄ selectivity in polymeric membranes from Morgan fingerprints. SHAP (SHapley Additive exPlanations) values are used to identify dominant structural descriptors and compare them with those governing CO₂ permeability.

Source: Zenodo (CERN European Organization for Nuclear Research)

DOI: 10.5281/zenodo.22075999

Author: Saki Norouzi, Ghazal

Language: English

Topic: Membrane Separation and Gas Transport

Subfield: Filtration and Separation

Field: Chemical Engineering

Domain: Physical Sciences

Open Access status: green

OpenAlex ID: W7204102193

Cites:

Cited by:

Related to: 10