You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Idea: create a new fixed cluster, cluster 1 that is initialised at the beginning with no spikes, and cannot be deleted (like the noise cluster 0). We force this cluster to have a diagonal covariance matrix which we compute explicitly. We don't compute the correction term. We do have a prior point but no noise point. No change to penalty computation.
This should remove some points that are causing large cluster masks and slowing down the computations.
The text was updated successfully, but these errors were encountered:
Idea: create a new fixed cluster, cluster 1 that is initialised at the beginning with no spikes, and cannot be deleted (like the noise cluster 0). We force this cluster to have a diagonal covariance matrix which we compute explicitly. We don't compute the correction term. We do have a prior point but no noise point. No change to penalty computation.
This should remove some points that are causing large cluster masks and slowing down the computations.
The text was updated successfully, but these errors were encountered: