search

lnccbrown / HSSM

Development of HSSM package
Other
70 stars 10 forks source link

suspicious upper bounds on t #447

Closed AlexanderFengler closed 2 weeks ago

AlexanderFengler commented 1 month ago

I ran into situations where the t parameter was upper-bounded by the lowest rt in the data, even though p_outlier > 0 was set. This shouldn't happen, and needs some investigation.

digicosmos86 commented 1 month ago

It turns out that we replace the likelihood with a lower_bound whenever rt < t regardless of the value of p_outlier. This should be relaxed when p_outlier > 0. A fix is coming in the next few days

AlexanderFengler commented 2 weeks ago

@digicosmos86 this is actually resolved with your latest PR right?