search

certifiedwaif / phd

5 stars 0 forks source link

Convergence seems sensitive to values of sigma2_u #4

Closed certifiedwaif closed 10 years ago

certifiedwaif commented 10 years ago

For example, testing with .5^2 works, whereas testing with .75^2 fails to converge.

JohnOrmerod commented 10 years ago

Dear Mark, Argh. Could be algorithm settings. Is it because NAs or Infs start showing up? John

From: Mark Greenaway [notifications@github.com] Sent: Monday, May 12, 2014 12:36 PM To: certifiedwaif/phd Subject: [phd] Convergence seems sensitive to values of sigma2_u (#4)

For example, testing with .5^2 works, whereas testing with .75^2 fails to converge.

— Reply to this email directly or view it on GitHubhttps://github.com/certifiedwaif/phd/issues/4.

certifiedwaif commented 10 years ago

That's right, -Inf and NaN. A test case which demonstrates this is expected_sigma2_u=.75^2

On Mon, May 12, 2014 at 12:48 PM, JohnOrmerod notifications@github.comwrote:

Dear Mark, Argh. Could be algorithm settings. Is it because NAs or Infs start showing up? John

From: Mark Greenaway [notifications@github.com] Sent: Monday, May 12, 2014 12:36 PM To: certifiedwaif/phd Subject: [phd] Convergence seems sensitive to values of sigma2_u (#4)

For example, testing with .5^2 works, whereas testing with .75^2 fails to converge.

— Reply to this email directly or view it on GitHub< https://github.com/certifiedwaif/phd/issues/4>.

— Reply to this email directly or view it on GitHubhttps://github.com/certifiedwaif/phd/issues/4#issuecomment-42792428 .