#StackBounty: #regression #scikit-learn #error #gaussian-process #uncertainty Accounting for errors in independent variable through Gau…

Bounty: 50

In Gaussian process regression (GPR), one applies a kernel (i.e. covariance function) to describe the similarity between observed and predicted data in the domain. The diagonal of the covariance function accounts for noise in the dependent or response variable. But is there an analogous method to errors-in-variables models (e.g. total least squares) to account for error in the independent or explanatory variable(s) inherently in GPR?


Get this bounty!!!

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.