Regression GWR depentent variable as a natural logarithm - estimate output ???

Discussion created by peter_m on Dec 17, 2012
Hallo Forum Members!

I made a hedonic house price model with the depentent variable price per square meter.

Now I have compute a model with the transformation of the natural logarithm from the price per square meter (LN).
- In a "ordinary" linear regression model (OLS) I get regression coefficents (estimators) in percentage (%)!

- Is it the same in the GWR ????

In the GWR model I get the following estimators:

For example does this mean:
year_ch1                    11%      
year_ch2                    16%     
year_ch3                    9%
year_ch4                    10%     
ant_ter                     31%   
z_akad                      33%
z_etw_i                     103%      
z_geb_9                    -18%   
z_laerm                    -46%    
pb_geb1                     31%     
p_r_v1                      98% 
pb_geb10                   -3%

But whats with the high min. max. Values!!! ??
z_etw_i                   -10.972751       18.846764

Thanks for your help!