|
Post by ks on Apr 22, 2014 2:04:42 GMT
Hi, Did you guys play around wiht other convergence criteria than the log-likelihood convergence mentioned in the paper?
Thanks
|
|
|
Post by Jian Yang on Apr 24, 2014 12:31:12 GMT
For example, squared difference in estimates of variance components between the current and previous iterations < an arbitrary small value.
|
|