DESCRIPTION

r.regression.line calculates a linear regression from two raster maps, according to the formula 
y = a + b*x
where 
x
y
represent the input raster maps.

Optionally, it saves regression coefficients as a ASCII file. The result includes the following coefficients: offset/intercept (a) and gain/slope (b), correlation coefficient (R), number of elements (N), means (medX, medY), standard deviations (sdX, sdY), and the F test for testing the significance of the regression model as a whole (F).

NOTES

The results for offset/intercept (a) and gain/slope (b) are identical to that obtained from R-stats's lm() function.

EXAMPLE

Comparison of two DEMs (SRTM and NED, both at 30m resolution), provided in the North Carolina sample dataset: 
g.region raster=elev_srtm_30m -p
r.regression.line mapx=elev_ned_30m mapy=elev_srtm_30m 
 y = a + b*x
   a (Offset): -1.659279
   b (Gain): 1.043968
   R (sumXY - sumX*sumY/N): 0.894038
   N (Number of elements): 225000
   F (F-test significance): 896093.366283
   meanX (Mean of map1): 110.307571
   sdX (Standard deviation of map1): 20.311998
   meanY (Mean of map2): 113.498292
   sdY (Standard deviation of map2): 23.718307

Using the script style flag AND eval to make results available in the shell: 

g.region raster=elev_srtm_30m -p
eval `r.regression.line -g mapx=elev_ned_30m mapy=elev_srtm_30m`

# print result stored in respective variables
echo $a
-1.659279

echo $b
1.043968

echo $R
0.894038

SEE ALSO

d.correlate, r.regression.multi, r.stats

AUTHORS

Dr. Agustin Lobo - alobo at ija.csic.es
Updated to GRASS 5.7 Michael Barton, Arizona State University
Script style output Markus Neteler
Conversion to C module Markus Metz

Last changed: $Date$