Closed jotaylor closed 6 years ago
Hi Jo, no it is not a bug. I purposely added that prior to help convergence. In essence it required that the sum of templates doesn't exceed the max of the data spectrum. You could remove it but the chains take way way longer to converge and logically we do not expect the sum of the templates to exceed the maximum so I don't see much harm in leaving it in (though could maybe add a buffer amount).
Hi Jo, no it is not a bug. I purposely added that prior to help convergence. In essence it required that the sum of templates doesn't exceed the max of the data spectrum. You could remove it but the chains take way way longer to converge and logically we do not expect the sum of the templates to exceed the maximum so I don't see much harm in leaving it in (though could maybe add a buffer amount).
From: Jo notifications@github.com<mailto:notifications@github.com> Reply-To: antking/SPAMM reply@reply.github.com<mailto:reply@reply.github.com> Date: Tuesday, 19 September 2017 5:44 am To: antking/SPAMM SPAMM@noreply.github.com<mailto:SPAMM@noreply.github.com> Cc: Anthea King anthea.king@uqconnect.edu.au<mailto:anthea.king@uqconnect.edu.au>, Assign assign@noreply.github.com<mailto:assign@noreply.github.com> Subject: [antking/SPAMM] Why is another prior added in HostGalaxyComponent? (#13) Resent-From: Anthea King anthea.king@uqconnect.edu.au<mailto:anthea.king@uqconnect.edu.au>
— You are receiving this because you were assigned. Reply to this email directly, view it on GitHubhttps://github.com/antking/SPAMM/issues/13, or mute the threadhttps://github.com/notifications/unsubscribe-auth/ADdqo6MstcLwtD-orYHKpFTlBxQGVQFAks5sjsgzgaJpZM4Pbbb1.
https://github.com/antking/SPAMM/blob/907abbe494c7fd45fc256c50712517c9943fc93c/spamm/components/HostGalaxyComponent.py#L322
In the WorkingHostGalaxyandProperlyRedshift branch another prior is added after normalization priors. Is this a bug?