Hello I am using the step function in order to do backward selection for a
linear model of 52 variables with the following commands:
object<-lm(vars[,1] ~ (vars[,2:(ncol(predictors)+1)]-1))
BackS<-step(object,direction="backward")
but it isn't dropping any if the variables in the model, but there are lots
of not significant variables as you can see here
> object<-lm(vars[,1] ~ (vars[,2:(ncol(predictors)+1)]-1))
> summary(object)
Call:
lm(formula = vars[, 1] ~ (vars[, 2:(ncol(predictors) + 1)] -
1))
Residuals:
Min 1Q Median 3Q Max
-0.56388 -0.10762 -0.01433 0.08495 0.82477
Coefficients:
Estimate Std.
Error t value Pr(>|t|)
vars[, 2:(ncol(predictors) + 1)]SS.1 0.028772
0.025458 1.130 0.260896
vars[, 2:(ncol(predictors) + 1)]Precio.Promedio.Bolsa[1] -0.308076
0.096243 -3.201 0.001795 **
vars[, 2:(ncol(predictors) + 1)]Precio.Promedio.Bolsa[2] 0.130134
0.101734 1.279 0.203559
vars[, 2:(ncol(predictors) + 1)]Precio.Promedio.Bolsa[3] 0.014345
0.106282 0.135 0.892887
vars[, 2:(ncol(predictors) + 1)]Precio.Promedio.Bolsa[4] -0.175958
0.107097 -1.643 0.103268
vars[, 2:(ncol(predictors) + 1)]Precio.Promedio.Bolsa[5] 0.016270
0.106081 0.153 0.878391
vars[, 2:(ncol(predictors) + 1)]Precio.Promedio.Bolsa[6] -0.089018
0.091132 -0.977 0.330834
vars[, 2:(ncol(predictors) + 1)]Precio.Promedio.Bolsa[7] -0.270550
0.075537 -3.582 0.000512 ***
vars[, 2:(ncol(predictors) + 1)]Precio.Promedio.Bolsa[8] -0.106691
0.074448 -1.433 0.154694
vars[, 2:(ncol(predictors) + 1)]Precio.Promedio.Bolsa[9] 0.118962
0.076886 1.547 0.124699
vars[, 2:(ncol(predictors) + 1)]Precio.Promedio.Bolsa[10] -0.055112
0.076225 -0.723 0.471218
vars[, 2:(ncol(predictors) + 1)]Precio.Promedio.Bolsa[11] -0.135113
0.076307 -1.771 0.079415 .
vars[, 2:(ncol(predictors) + 1)]Precio.Promedio.Bolsa[12] 0.082478
0.075130 1.098 0.274707
vars[, 2:(ncol(predictors) + 1)]Anomalia[0] 0.123054
0.213980 0.575 0.566426
vars[, 2:(ncol(predictors) + 1)]Anomalia[1] 0.078511
0.507544 0.155 0.877353
vars[, 2:(ncol(predictors) + 1)]Anomalia[2] -0.399726
0.581594 -0.687 0.493357
vars[, 2:(ncol(predictors) + 1)]Anomalia[3] -0.002103
0.583109 -0.004 0.997129
vars[, 2:(ncol(predictors) + 1)]Anomalia[4] 0.596937
0.678115 0.880 0.380640
vars[, 2:(ncol(predictors) + 1)]Anomalia[5] -0.547555
0.710687 -0.770 0.442695
vars[, 2:(ncol(predictors) + 1)]Anomalia[6] -0.142452
0.678536 -0.210 0.834106
vars[, 2:(ncol(predictors) + 1)]Anomalia[7] 0.506431
0.692960 0.731 0.466455
vars[, 2:(ncol(predictors) + 1)]Anomalia[8] -0.117177
0.662596 -0.177 0.859958
vars[, 2:(ncol(predictors) + 1)]Anomalia[9] -0.550570
0.563421 -0.977 0.330638
vars[, 2:(ncol(predictors) + 1)]Anomalia[10] 0.799499
0.555007 1.441 0.152587
vars[, 2:(ncol(predictors) + 1)]Anomalia[11] -0.577416
0.504046 -1.146 0.254485
vars[, 2:(ncol(predictors) + 1)]Anomalia[12] 0.204479
0.221030 0.925 0.356948
vars[, 2:(ncol(predictors) + 1)]demanda.nacional[0] -0.572351
1.303885 -0.439 0.661561
vars[, 2:(ncol(predictors) + 1)]demanda.nacional[1] 0.270387
1.715912 0.158 0.875082
vars[, 2:(ncol(predictors) + 1)]demanda.nacional[2] 1.939207
1.806931 1.073 0.285549
vars[, 2:(ncol(predictors) + 1)]demanda.nacional[3] 1.501964
1.779253 0.844 0.400432
vars[, 2:(ncol(predictors) + 1)]demanda.nacional[4] 1.292790
1.759802 0.735 0.464147
vars[, 2:(ncol(predictors) + 1)]demanda.nacional[5] 1.197978
1.760600 0.680 0.497670
vars[, 2:(ncol(predictors) + 1)]demanda.nacional[6] 0.338068
1.720709 0.196 0.844608
vars[, 2:(ncol(predictors) + 1)]demanda.nacional[7] -2.197186
1.616212 -1.359 0.176805
vars[, 2:(ncol(predictors) + 1)]demanda.nacional[8] -2.050263
1.542936 -1.329 0.186687
vars[, 2:(ncol(predictors) + 1)]demanda.nacional[9] -0.103823
1.541956 -0.067 0.946441
vars[, 2:(ncol(predictors) + 1)]demanda.nacional[10] 0.349220
1.545823 0.226 0.821693
vars[, 2:(ncol(predictors) + 1)]demanda.nacional[11] -0.654607
1.476141 -0.443 0.658313
vars[, 2:(ncol(predictors) + 1)]demanda.nacional[12] -0.254144
1.193506 -0.213 0.831772
vars[, 2:(ncol(predictors) + 1)]Nivel.Embalse[0] -1.500119
0.428395 -3.502 0.000671 ***
vars[, 2:(ncol(predictors) + 1)]Nivel.Embalse[1] -1.058775
0.475011 -2.229 0.027869 *
vars[, 2:(ncol(predictors) + 1)]Nivel.Embalse[2] 0.818735
0.497920 1.644 0.102994
vars[, 2:(ncol(predictors) + 1)]Nivel.Embalse[3] 0.057331
0.528216 0.109 0.913769
vars[, 2:(ncol(predictors) + 1)]Nivel.Embalse[4] -0.529271
0.519284 -1.019 0.310350
vars[, 2:(ncol(predictors) + 1)]Nivel.Embalse[5] -0.649193
0.508210 -1.277 0.204171
vars[, 2:(ncol(predictors) + 1)]Nivel.Embalse[6] 0.511649
0.490911 1.042 0.299605
vars[, 2:(ncol(predictors) + 1)]Nivel.Embalse[7] -0.545404
0.473994 -1.151 0.252392
vars[, 2:(ncol(predictors) + 1)]Nivel.Embalse[8] -0.314593
0.489687 -0.642 0.521939
vars[, 2:(ncol(predictors) + 1)]Nivel.Embalse[9] -0.091112
0.510613 -0.178 0.858712
vars[, 2:(ncol(predictors) + 1)]Nivel.Embalse[10] -0.030684
0.492553 -0.062 0.950442
vars[, 2:(ncol(predictors) + 1)]Nivel.Embalse[11] 0.162751
0.488237 0.333 0.739515
vars[, 2:(ncol(predictors) + 1)]Nivel.Embalse[12] 0.370126
0.458473 0.807 0.421250
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.2159 on 109 degrees of freedom
Multiple R-squared: 0.566, Adjusted R-squared: 0.359
F-statistic: 2.734 on 52 and 109 DF, p-value: 5.24e-06
do you know how can I do this? how can I do backward selection on a
regression without an intercept? Thank you
Felipe Parra
[[alternative HTML version deleted]]
