Goodness of Fit

Computes goodness of fit (model performance) for regression.

Description

Computes goodness of fit for regression model given vector of target and output values.

USAGE:
   [gf] = gfit(t,y)
   [gf] = gfit(t,y,gFitMeasure)
     
INPUT
t: vector of target values for regression model

y: vector of output from regression model.

gFitMeasure: string value representing different form of goodness of fit measure (for eg. '1' - mean squarred error, '5' - mean absolute error)
         
OUTPUT:
 gf: goodness of fit values between model output and target

• Acknowledgements: This submission has inspired the following:; Goodness of Fit (Modified)
• MATLAB release: MATLAB 7.0.1 (R14SP1)

[al azhar]

just what i need

[Felix Hebeler]

Script does not check for NaNs, and matrices have to be reshaped. If you need a simple script to calculate RMSE, that considers NaNs and operates on matrices, please check
http://www.mathworks.com/matlabcentral/fileexchange/loadFile.do?objectId=21383&objectType=FILE

[Clement Atzberger]

Function file performs very well. Good work! Unfortunately, only vectors can be input. Also I miss the possibility to calculate several statistics at the same function call. Please check your expression 'squarred' (sic!) misspelled throughout the file. This is correctly spelled as 'squared' ... with one 'r' ;-)