Finding the best not the most: regularized loss minimization subgraph selection for graph classification
作者：
Highlights：
• Our algorithm selects the optimal subgaphs for graph classification.
• We generalize the column generation technique of gBoost for graph classification.
• We use the elastic net to produce sparse and robust solutions for subgraph selection.
• We derive an effective pruning rule for search space reduction.
• We demonstrate the effectiveness of our algorithm.
摘要
Highlights•Our algorithm selects the optimal subgaphs for graph classification.•We generalize the column generation technique of gBoost for graph classification.•We use the elastic net to produce sparse and robust solutions for subgraph selection.•We derive an effective pruning rule for search space reduction.•We demonstrate the effectiveness of our algorithm.
论文关键词：Feature selection,Classification,Graph classification,Sparse learning
论文评审过程：Received 18 August 2014, Revised 15 May 2015, Accepted 18 May 2015, Available online 31 May 2015, Version of Record 16 July 2015.
论文官网地址：https://doi.org/10.1016/j.patcog.2015.05.019
