[R-sig-Geo] [SpatialFiltering] constant std error for eigenvectors in the linear model

[Aquarius]

Hi y'all,

I did a simple Spatial Filtering OLS model on land value with explanatory
variables: distance to CBD, slope, and other covaraites. The line of code is
copied below:

results= SpatialFiltering(LandValue_08 ~
Dist2CBD05+distfwy+distmaj+distmin+Slope+periarea, data=lu2,
nb=tmp$neighbours, style="W", ExactEV=TRUE, zero.policy=TRUE)

Then I used the selected 12 eigenvectors in the linear model:

results1=lm(LandValue_08 ~
Dist2CBD05+distfwy+distmaj+distmin+Slope+periarea+fitted(results),
data=lu2);

However, the std. error for the 12 eigenvectors (which serve as "synthetic
variables") are constant with value 9.254, as shown below:

 	Estimate	Stdev	T-Stats	Pr(>|t|)
Constant	24.09	0.799	30.158	< 2e-16
DistCBD	-1.747	0.104	-16.84	< 2e-16
DistFwy	2.221	0.229	9.696	< 2e-16
DistMaj	2.051	0.776	2.642	0.00846
DistMin	-0.8192	0.347	-2.359	0.01866
Slope	-0.2761	0.136	-2.037	0.04209
vec5	-176.8	9.254	-19.1	< 2e-16
vec3	-58.99	9.254	-6.374	3.86E-10
vec4	-53.27	9.254	-5.756	1.42E-08
vec2	49.86	9.254	5.387	1.06E-07
vec14	37.54	9.254	4.057	5.68E-05
vec31	37.50	9.254	4.052	5.79E-05
vec48	38.77	9.254	4.189	3.25E-05
vec57	34.17	9.254	3.692	0.00024
vec19	26.24	9.254	2.836	0.00474
vec33	26.92	9.254	2.909	0.00377
vec61	-32.43	9.254	-3.504	0.00049
vec6	21.04	9.254	2.274	0.02335
vec7	20.49	9.254	2.214	0.02724
vec16	20.90	9.254	2.258	0.02431
vec15	20.48	9.254	2.213	0.02728
vec46	23.87	9.254	2.579	0.01016
vec23	20.47	9.254	2.212	0.02734
vec40	-22.29	9.254	-2.409	0.01634
vec75	28.06	9.254	3.032	0.00254
vec18	18.81	9.254	2.032	0.04259
vec39	20.71	9.254	2.238	0.02561
vec84	28.32	9.254	3.06	0.00232
vec66	-23.46	9.254	-2.535	0.01153
vec86	26.56	9.254	2.87	0.00426
vec92	25.73	9.254	2.78	0.00562
vec12	15.40	9.254	1.664	0.09666


What is the cause for the constant std error?

Thanks a million in advance :)

Yiyi

--
View this message in context: http://r-sig-geo.2731867.n2.nabble.com/SpatialFiltering-constant-std-error-for-eigenvectors-in-the-linear-model-tp6668793p6668793.html
Sent from the R-sig-geo mailing list archive at Nabble.com.