Rmse

From Eigenvector Research Documentation Wiki
Revision as of 10:00, 9 October 2008 by imported>Chuck
Jump to navigation Jump to search

Purpose

Calculate Root Mean Square Difference (Error).

Synopsis

err = rmse(y1,y2)

Description

RMSE is used to calculate the root mean square difference between two vectors or matrices. If the vectors or matrices contain matching model-estimated and measured quantities, then the output is the Root Mean Square Error (RMSE).

The output err depends on the input, and there are 3 possible cases, outlined below:

Input and Output Cases

  • Case A): y1 is a matrix or vector
err = rmse(y1);
The output err is the root mean square of the elements of y1.
  • Case B): y1 is a matrix or vector, y2 is the same size as y1
err = rmse(y1,y2);
The output err is the root mean square of the difference between y1 and y2.
  • Case C): y1 is a matrix or vector, y2 is a column vector
err = rmse(y1,y2);
The output err is the root mean square of the difference between each column of y1 and y2.
Example: For example, y2 is a reference and the rmse is calculated between each column of y1 and the vector y2.

See Also

crossval