Multiple Linear Regression with Fit and Cross Validation Statistics

What it does

The function mlr.m generates an MLR model fit and does `leave one out' cross-validation of the model. Measures of R-squared, Adjusted R-squared, root-mean-square error of calibration (RMSEC), and root-mean-square error of cross-validation (RMSECV) are printed to the screen. The function also generates a table with the true and fitted values are along with the relative percent error (RPE) and upper and lower limits of an approximate 95% confidence interval on future observed values assuming normality. An example "session" with mlr.m taken from the MATLAB command window is shown below.

The mlr.m function generates several plots, such as the actual versus fitted values with error bounds (shown below) and actual vs. leave one out calibration predictions.

 

Requirements for running corrmap

• MATLAB 4.2 or 5.0
• No other toolboxes required

Developed by:

Kirk Remund
Pete Eschbach
Pacific Northwest National Laboratory - Battelle

e-mail: km_remund@pnl.gov
      : pa_eschbach@pnl.gov
phone :  (509) 372-4729
               375-2678 (pete)
fax:     (509) 375-3614
               372-4725 (pete)

Download mlr.m

To get it, simply click on mlr.m for Mac, or mlr.m for PC. Move the file to a folder on your MATLAB path and you're done.

Return to MATLAB User Area.