Negative data leading to std < 0

DOI-USGS / geobipy

Geophysical Bayesian Inference in Python. Docs:

https://usgs.github.io/geobipy/

Other

78 stars 19 forks source link

Negative data leading to std < 0 #14

Closed bminsley closed 5 years ago

bminsley commented 5 years ago

changed error calculation to use absolute values of data in order to account for negative data that produce std < 0. Should still warn user in .out file if data contain negative values larger in magnitude than -10 or so .. these may be related to susceptibility, not just low-amplitude noise

leonfoks commented 5 years ago

The absolute value here doesn't affect the final std because we square each item before summation. Ill dig a little further and figure out where this is happening

bminsley commented 5 years ago

Hmm good point about squaring values...