Can correlation matrix be made visible for Levenberg Marquardt? (Trac #911)

[RichardHeenan]

Can the parameter correlation matrix be made visible at all for Levenberg Marquardt?

(The L-M method modifies the least squares matrix, so may require some care o get the desired result.)

Migrated from http://trac.sasview.org/ticket/911

{
    "status": "new",
    "changetime": "2017-10-23T14:25:01",
    "_ts": "2017-10-23 14:25:01.882834+00:00",
    "description": "Can the parameter correlation matrix be made visible at all for Levenberg Marquardt?\n\n(The L-M method modifies the least squares matrix, so may require some care o get the desired result.) ",
    "reporter": "richardh",
    "cc": "",
    "resolution": "",
    "workpackage": "SasView Framework Enhancements",
    "time": "2017-04-04T08:19:42",
    "component": "SasView",
    "summary": "Can correlation matrix be made visible for Levenberg Marquardt?",
    "priority": "trivial",
    "keywords": "",
    "milestone": "SasView 5.0.0",
    "owner": "pkienzle",
    "type": "enhancement"
}

[pkienzle]

Trac update at 2017/04/11 10:12:30: pkienzle commented:

sas/sascalc/fit/BumpsFitting.py, line 290, can call fitness.cov() to get the covariance matrix from bumps. This would need to be returned to the user interface and a new output panel created for displaying it.

[butlerpd]

Trac update at 2017/10/23 14:25:01: butler changed milestone from "SasView 4.2.0" to "SasView 5.0.0"