Comments Regarding `On the Identifiability of the Influence Model for Stochastic Spatiotemporal Spread Processes'
This is a critique of a specific prior work, pointing out technical errors and omissions, which is incremental and relevant only to researchers working on identifiability of networked Markov chain models.
The paper identifies two errors in a prior identifiability analysis of the influence model for stochastic spatiotemporal spread processes, specifically regarding unidentifiability under partial observation and a missing recurrence criterion, and raises concerns about the problem formulation and estimation approach.
The identifiability analysis of a networked Markov chain model known as the influence model, as described in a recent contribution to Arxiv, is examined. Two errors in the identifiability analysis -- one related to the unidentifiability of the partially-observed influence model, the second related to an omission of an additional recurrence criterion for identifiability -- are noted. In addition, some concerns about the formulation of the identifiability problem and the proposed estimation approach are noted.