Tag
The paper proposes MambaLSTM, a framework combining Mamba state-space models and LSTM for spatio-temporal traffic accident risk prediction, addressing noise in feature fusion and global spatial correlation.