Simple beats deep: linear interpolation outperforms AI models when real-world clinical data goes missing
A new benchmark study shows that simple linear imputation methods outperform deep learning models when realistic spatio-temporal missingness is simulated ...
