channel-closures

This paper explores predicting whether Lightning Network channels will close mutually or via forced closure using machine learning on gossip data. An MLP with temporal features outperforms graph-based models, and the dataset is publicly released.