the previous issue was that we included the probability of an infection being observable in the future. This is no longer the case: by cutting off the future portions of the convolution matrix, the colsums now only sum the probabilities of being observed in each day within the target dates range
the previous issue was that we included the probability of an infection being observable in the future. This is no longer the case: by cutting off the future portions of the convolution matrix, the colsums now only sum the probabilities of being observed in each day within the target dates range