the baseline solution M6 is modified such that instead of 15 hand-crafted features, the neural network is instead trained on features extracted from a graph embedding. We use two different embedding approaches. The first method is employs node2vec (M7A)45, for which...
This is because we found that the distribution of the best-fit residuals of the cell fraction data resembles a Gaussian while that for the residuals of the logit-transformed cell fraction data does not (not shown). The nonlinear regression was performed using successive applications of python's ...