HawkesSumGaussian seems not to properly implement the method described in the cited paper.
After attempting to use tick for a while and failing to get good results from the HawkesSumGaussian method, I tried using the reference implementation of this function (written in Matlab) instead. I found that the reference implementation gives excellent results, and the tick implementation does not.
I'm not sure what errors were made in implementation, but one huge red flag is identified in #467. I dont know if this is the only problem, however.
Hi all,
HawkesSumGaussian seems not to properly implement the method described in the cited paper.
After attempting to use tick for a while and failing to get good results from the HawkesSumGaussian method, I tried using the reference implementation of this function (written in Matlab) instead. I found that the reference implementation gives excellent results, and the tick implementation does not.
I'm not sure what errors were made in implementation, but one huge red flag is identified in #467. I dont know if this is the only problem, however.