Researchers have developed path_boost, a new Python package designed for interpretable supervised learning on graph-structured data. The package implements PathBoost, a gradient boosting algorithm that identifies predictive labeled paths within graphs, offering greater interpretability than typical graph neural networks. PathBoost works by iteratively selecting and extending paths based on their predictive power, creating an additive model that clearly shows which substructures influence predictions. The package is compatible with scikit-learn, supports custom base learners, and includes features like automatic starting node selection and variable importance computation. It has been demonstrated on molecular property prediction and benchmarked against graph neural networks and graph kernel methods. AI
IMPACT Provides a more interpretable alternative to graph neural networks for structured data analysis.
RANK_REASON The cluster describes a new academic paper introducing a novel software package for graph-level prediction. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
