We should consider returning the probability mass (i.e., values on p and r vector) from unavailable nodes to the seeds (on r vector). That is, if current_id is not available upon API query, we should put do the following two steps:
(1) r[seednode_id] = r[seednode_id] + r[current_id] + p[current_id]
(2) remove current_id from everywhere
We should consider returning the probability mass (i.e., values on
p
andr
vector) from unavailable nodes to the seeds (onr
vector). That is, ifcurrent_id
is not available upon API query, we should put do the following two steps: (1)r[seednode_id] = r[seednode_id] + r[current_id] + p[current_id]
(2) remove current_id from everywhere