-
Notifications
You must be signed in to change notification settings - Fork 16
Regression
Thomas Nipen edited this page Aug 29, 2018
·
4 revisions
This scheme applies a linear regression to the variable.
The order is determined by the number of regression parameters.
- order (integer): A value of 0 uses the equation (y = a), a value of 1 (y = a + b * x), a value of 2 (y = a + b * x + c * x ^ 2)
- variables (string): If specified, use this list of variables as the predictands in a multivariate regression. For example if variables=precipitation,humidity then the following equation is used: a + b * precipitation + c * humidity. In this case precipitation and humidity fields must be available.
Global or local. The nearest parameter set is used for local parameter files. The number of parameters must be 1 more than the regression order. For multi-variate regression, the number of parameters must be 1 more than the number of variables.