search

lasiadhi / USACE-Bathymetry

0 stars 2 forks source link

Talk to Lasith's wife about sigma #87

Closed gmuldoon closed 8 years ago

gmuldoon commented 8 years ago

She's amazing

gmuldoon commented 8 years ago

sigma has to be outside the sum in the bayesian likelihood formulation because all errors are assumed to come from the same normal distribution.

conclusion: sigma=mean of sigma_{k,x}.

gmuldoon commented 8 years ago

log likelihood = - sum[(k_m - k_d)^2]/(2sig^2)