Build regression model from a set of candidate predictor variables by entering and removing predictors based on p values, in a stepwise manner until there is no variable left to enter or remove any more.
Usage
ols_step_both_p(model, ...)
# Default S3 method
ols_step_both_p(
model,
p_enter = 0.1,
p_remove = 0.3,
include = NULL,
exclude = NULL,
progress = FALSE,
details = FALSE,
steps = NULL,
...
)
# S3 method for class 'ols_step_both_p'
plot(x, model = NA, print_plot = TRUE, details = TRUE, ...)Arguments
- model
An object of class
lm; the model should include all candidate predictor variables.- ...
Other arguments.
- p_enter
p value; variables with p value less than
p_enterwill enter into the model.- p_remove
p value; variables with p more than
p_removewill be removed from the model.- include
Character or numeric vector; variables to be included in selection process.
- exclude
Character or numeric vector; variables to be excluded from selection process.
- progress
Logical; if
TRUE, will display variable selection progress.- details
Logical; if
TRUE, will print the regression result at each step.- steps
Number of steps after which the stepwise procedures should stop.
- x
An object of class
ols_step_both_p.- print_plot
logical; if
TRUE, prints the plot else returns a plot object.
Value
ols_step_both_p returns an object of class "ols_step_both_p".
An object of class "ols_step_both_p" is a list containing the
following components:
- model
final model; an object of class
lm- metrics
selection metrics
- beta_pval
beta and p values of models in each selection step
Details
Refits accept explicit contrast matrices or the base contrast names
contr.treatment, contr.sum, contr.helmert, contr.poly and contr.SAS
(resolved from the stats namespace); models fitted with custom named contrast
functions must be refitted with an explicit matrix, for example
contrasts = list(g = my_matrix). Refits carry the fitted model's frozen total
offset as an argument with all formula offset() terms removed for
standardized-beta and selection computations. Returned stepwise models are
fitted from the original call, preserving formula and argument offsets for
prediction on new data. Their coefficients, fitted values and total offset
must be identical to the frozen-offset selection fit; otherwise selection
stops with an error because prediction with new data is not supported for
that model.
For non-hierarchical stepwise selection, candidate rankings and criterion improvement tests use 10 significant digits: values that agree to 10 significant digits are ties, equal criterion values mean no improvement, and a remaining tie goes to the first candidate in formula order after any F tie is broken by p value (smaller for entry, larger for removal). P-value entry and removal cutoffs are compared without rounding.
References
Chatterjee, Samprit and Hadi, Ali. Regression Analysis by Example. 5th ed. N.p.: John Wiley & Sons, 2012. Print.
Examples
# \donttest{
# stepwise regression
model <- lm(y ~ ., data = surgical)
ols_step_both_p(model)
#>
#>
#> Stepwise Summary
#> ------------------------------------------------------------------------------
#> Step Variable AIC SBC SBIC R2 Adj. R2
#> ------------------------------------------------------------------------------
#> 0 Base Model 802.606 806.584 646.794 0.00000 0.00000
#> 1 liver_test (+) 771.875 777.842 616.009 0.45454 0.44405
#> 2 alc_heavy (+) 761.439 769.395 605.506 0.56674 0.54975
#> 3 enzyme_test (+) 750.509 760.454 595.297 0.65900 0.63854
#> 4 pindex (+) 735.715 747.649 582.943 0.75015 0.72975
#> 5 bcs (+) 730.620 744.543 579.638 0.78091 0.75808
#> ------------------------------------------------------------------------------
#>
#> Final Model Output
#> ------------------
#>
#> Model Summary
#> -------------------------------------------------------------------
#> R 0.884 RMSE 195.454
#> R-Squared 0.781 MSE 38202.426
#> Adj. R-Squared 0.758 Coef. Var 27.839
#> Pred R-Squared 0.700 AIC 730.620
#> MAE 137.656 SBC 744.543
#> -------------------------------------------------------------------
#> RMSE: Root Mean Square Error
#> MSE: Mean Square Error
#> MAE: Mean Absolute Error
#> AIC: Akaike Information Criteria
#> SBC: Schwarz Bayesian Criteria
#>
#> ANOVA
#> -----------------------------------------------------------------------
#> Sum of
#> Squares DF Mean Square F Sig.
#> -----------------------------------------------------------------------
#> Regression 6535804.090 5 1307160.818 34.217 0.0000
#> Residual 1833716.447 48 38202.426
#> Total 8369520.537 53
#> -----------------------------------------------------------------------
#>
#> Parameter Estimates
#> ------------------------------------------------------------------------------------------------
#> model Beta Std. Error Std. Beta t Sig lower upper
#> ------------------------------------------------------------------------------------------------
#> (Intercept) -1178.330 208.682 -5.647 0.000 -1597.914 -758.746
#> liver_test 58.064 40.144 0.156 1.446 0.155 -22.652 138.779
#> alc_heavy 317.848 71.634 0.314 4.437 0.000 173.818 461.878
#> enzyme_test 9.748 1.656 0.521 5.887 0.000 6.419 13.077
#> pindex 8.924 1.808 0.380 4.935 0.000 5.288 12.559
#> bcs 59.864 23.060 0.241 2.596 0.012 13.498 106.230
#> ------------------------------------------------------------------------------------------------
#>
# stepwise regression plot
model <- lm(y ~ ., data = surgical)
k <- ols_step_both_p(model)
plot(k)
# selection metrics
k$metrics
#> step variable method r2 adj_r2 aic sbc sbic
#> 1 1 liver_test addition 0.4545389 0.4440492 771.8753 777.8423 616.0089
#> 2 2 alc_heavy addition 0.5667409 0.5497504 761.4394 769.3953 605.5062
#> 3 3 enzyme_test addition 0.6590000 0.6385400 750.5089 760.4538 595.2974
#> 4 4 pindex addition 0.7501457 0.7297495 735.7146 747.6485 582.9426
#> 5 5 bcs addition 0.7809054 0.7580831 730.6204 744.5433 579.6377
#> mallows_cp rmse
#> 1 62.511923 296.2992
#> 2 41.368078 266.6484
#> 3 24.337853 238.9145
#> 4 7.537284 206.5835
#> 5 3.192498 195.4544
# final model
k$model
#>
#> Call:
#> (function (formula, data, subset, weights, na.action, method = "qr",
#> model = TRUE, x = FALSE, y = FALSE, qr = TRUE, singular.ok = TRUE,
#> contrasts = NULL, offset, ...)
#> {
#> ret.x <- x
#> ret.y <- y
#> cl <- match.call()
#> mf <- match.call(expand.dots = FALSE)
#> m <- match(c("formula", "data", "subset", "weights", "na.action",
#> "offset"), names(mf), 0L)
#> mf <- mf[c(1L, m)]
#> mf$drop.unused.levels <- TRUE
#> mf[[1L]] <- quote(stats::model.frame)
#> mf <- eval(mf, parent.frame())
#> if (method == "model.frame")
#> return(mf)
#> else if (method != "qr")
#> warning(gettextf("method = '%s' is not supported. Using 'qr'",
#> method), domain = NA)
#> mt <- attr(mf, "terms")
#> y <- model.response(mf, "numeric")
#> w <- as.vector(model.weights(mf))
#> if (!is.null(w) && !is.numeric(w))
#> stop("'weights' must be a numeric vector")
#> offset <- model.offset(mf)
#> mlm <- is.matrix(y)
#> ny <- if (mlm)
#> nrow(y)
#> else length(y)
#> if (!is.null(offset)) {
#> if (!mlm)
#> offset <- as.vector(offset)
#> if (NROW(offset) != ny)
#> stop(gettextf("number of offsets is %d, should equal %d (number of observations)",
#> NROW(offset), ny), domain = NA)
#> }
#> if (is.empty.model(mt)) {
#> x <- NULL
#> z <- list(coefficients = if (mlm) matrix(NA_real_, 0,
#> ncol(y)) else numeric(), residuals = y, fitted.values = 0 *
#> y, weights = w, rank = 0L, df.residual = if (!is.null(w)) sum(w !=
#> 0) else ny)
#> if (!is.null(offset)) {
#> z$fitted.values <- offset
#> z$residuals <- y - offset
#> }
#> }
#> else {
#> x <- model.matrix(mt, mf, contrasts)
#> z <- if (is.null(w))
#> lm.fit(x, y, offset = offset, singular.ok = singular.ok,
#> ...)
#> else lm.wfit(x, y, w, offset = offset, singular.ok = singular.ok,
#> ...)
#> }
#> class(z) <- c(if (mlm) "mlm", "lm")
#> z$na.action <- attr(mf, "na.action")
#> z$offset <- offset
#> z$contrasts <- attr(x, "contrasts")
#> z$xlevels <- .getXlevels(mt, mf)
#> z$call <- cl
#> z$terms <- mt
#> if (model)
#> z$model <- mf
#> if (ret.x)
#> z$x <- x
#> if (ret.y)
#> z$y <- y
#> if (!qr)
#> z$qr <- NULL
#> z
#> })(formula = y ~ liver_test + alc_heavy + enzyme_test + pindex +
#> bcs, data = surgical)
#>
#> Coefficients:
#> (Intercept) liver_test alc_heavy enzyme_test pindex bcs
#> -1178.330 58.064 317.848 9.748 8.924 59.864
#>
# include or exclude variables
model <- lm(y ~ ., data = stepdata)
# force variable to be included in selection process
ols_step_both_p(model, include = c("x6"))
#>
#>
#> Stepwise Summary
#> --------------------------------------------------------------------------------
#> Step Variable AIC SBC SBIC R2 Adj. R2
#> --------------------------------------------------------------------------------
#> 0 Base Model 33473.297 33497.007 -23285.069 0.69145 0.69143
#> 1 x1 (+) 32931.758 32963.372 -23826.833 0.69972 0.69969
#> 2 x3 (+) 31912.722 31952.239 -24845.827 0.71466 0.71462
#> 3 x2 (+) 29304.296 29351.717 -27453.243 0.74958 0.74953
#> 4 x4 (+) 29302.797 29358.121 -27454.740 0.74962 0.74956
#> --------------------------------------------------------------------------------
#>
#> Final Model Output
#> ------------------
#>
#> Model Summary
#> -----------------------------------------------------------------
#> R 0.866 RMSE 0.503
#> R-Squared 0.750 MSE 0.253
#> Adj. R-Squared 0.750 Coef. Var 6431.017
#> Pred R-Squared 0.749 AIC 29302.797
#> MAE 0.402 SBC 29358.121
#> -----------------------------------------------------------------
#> RMSE: Root Mean Square Error
#> MSE: Mean Square Error
#> MAE: Mean Absolute Error
#> AIC: Akaike Information Criteria
#> SBC: Schwarz Bayesian Criteria
#>
#> ANOVA
#> ---------------------------------------------------------------------------
#> Sum of
#> Squares DF Mean Square F Sig.
#> ---------------------------------------------------------------------------
#> Regression 15163.532 5 3032.706 11972.264 0.0000
#> Residual 5064.701 19994 0.253
#> Total 20228.233 19999
#> ---------------------------------------------------------------------------
#>
#> Parameter Estimates
#> ---------------------------------------------------------------------------------------
#> model Beta Std. Error Std. Beta t Sig lower upper
#> ---------------------------------------------------------------------------------------
#> (Intercept) -0.005 0.004 -1.497 0.135 -0.012 0.002
#> x6 0.000 0.004 -0.002 -0.129 0.897 -0.007 0.007
#> x1 0.256 0.005 0.363 54.592 0.000 0.247 0.265
#> x3 0.253 0.005 0.356 53.998 0.000 0.244 0.262
#> x2 0.249 0.005 0.347 52.769 0.000 0.240 0.258
#> x4 -0.007 0.004 -0.007 -1.870 0.061 -0.014 0.000
#> ---------------------------------------------------------------------------------------
#>
# use index of variable instead of name
ols_step_both_p(model, include = c(6))
#>
#>
#> Stepwise Summary
#> --------------------------------------------------------------------------------
#> Step Variable AIC SBC SBIC R2 Adj. R2
#> --------------------------------------------------------------------------------
#> 0 Base Model 33473.297 33497.007 -23285.069 0.69145 0.69143
#> 1 x1 (+) 32931.758 32963.372 -23826.833 0.69972 0.69969
#> 2 x3 (+) 31912.722 31952.239 -24845.827 0.71466 0.71462
#> 3 x2 (+) 29304.296 29351.717 -27453.243 0.74958 0.74953
#> 4 x4 (+) 29302.797 29358.121 -27454.740 0.74962 0.74956
#> --------------------------------------------------------------------------------
#>
#> Final Model Output
#> ------------------
#>
#> Model Summary
#> -----------------------------------------------------------------
#> R 0.866 RMSE 0.503
#> R-Squared 0.750 MSE 0.253
#> Adj. R-Squared 0.750 Coef. Var 6431.017
#> Pred R-Squared 0.749 AIC 29302.797
#> MAE 0.402 SBC 29358.121
#> -----------------------------------------------------------------
#> RMSE: Root Mean Square Error
#> MSE: Mean Square Error
#> MAE: Mean Absolute Error
#> AIC: Akaike Information Criteria
#> SBC: Schwarz Bayesian Criteria
#>
#> ANOVA
#> ---------------------------------------------------------------------------
#> Sum of
#> Squares DF Mean Square F Sig.
#> ---------------------------------------------------------------------------
#> Regression 15163.532 5 3032.706 11972.264 0.0000
#> Residual 5064.701 19994 0.253
#> Total 20228.233 19999
#> ---------------------------------------------------------------------------
#>
#> Parameter Estimates
#> ---------------------------------------------------------------------------------------
#> model Beta Std. Error Std. Beta t Sig lower upper
#> ---------------------------------------------------------------------------------------
#> (Intercept) -0.005 0.004 -1.497 0.135 -0.012 0.002
#> x6 0.000 0.004 -0.002 -0.129 0.897 -0.007 0.007
#> x1 0.256 0.005 0.363 54.592 0.000 0.247 0.265
#> x3 0.253 0.005 0.356 53.998 0.000 0.244 0.262
#> x2 0.249 0.005 0.347 52.769 0.000 0.240 0.258
#> x4 -0.007 0.004 -0.007 -1.870 0.061 -0.014 0.000
#> ---------------------------------------------------------------------------------------
#>
# force variable to be excluded from selection process
ols_step_both_p(model, exclude = c("x1"))
#>
#>
#> Stepwise Summary
#> --------------------------------------------------------------------------------
#> Step Variable AIC SBC SBIC R2 Adj. R2
#> --------------------------------------------------------------------------------
#> 0 Base Model 56988.482 57004.289 228.318 0.00000 0.00000
#> 1 x6 (+) 33473.297 33497.007 -23285.069 0.69145 0.69143
#> 2 x3 (+) 32959.771 32991.385 -23798.829 0.69930 0.69927
#> 3 x2 (+) 32080.424 32119.941 -24678.188 0.71226 0.71222
#> 4 x4 (+) 32079.710 32127.131 -24679.159 0.71230 0.71224
#> --------------------------------------------------------------------------------
#>
#> Final Model Output
#> ------------------
#>
#> Model Summary
#> -----------------------------------------------------------------
#> R 0.844 RMSE 0.539
#> R-Squared 0.712 MSE 0.291
#> Adj. R-Squared 0.712 Coef. Var 6893.511
#> Pred R-Squared 0.712 AIC 32079.710
#> MAE 0.430 SBC 32127.131
#> -----------------------------------------------------------------
#> RMSE: Root Mean Square Error
#> MSE: Mean Square Error
#> MAE: Mean Absolute Error
#> AIC: Akaike Information Criteria
#> SBC: Schwarz Bayesian Criteria
#>
#> ANOVA
#> ---------------------------------------------------------------------------
#> Sum of
#> Squares DF Mean Square F Sig.
#> ---------------------------------------------------------------------------
#> Regression 14408.579 4 3602.145 12376.147 0.0000
#> Residual 5819.654 19995 0.291
#> Total 20228.233 19999
#> ---------------------------------------------------------------------------
#>
#> Parameter Estimates
#> ---------------------------------------------------------------------------------------
#> model Beta Std. Error Std. Beta t Sig lower upper
#> ---------------------------------------------------------------------------------------
#> (Intercept) -0.004 0.004 -1.006 0.315 -0.011 0.004
#> x6 0.148 0.002 0.528 59.728 0.000 0.143 0.153
#> x3 0.141 0.005 0.199 31.259 0.000 0.132 0.150
#> x2 0.137 0.005 0.190 29.983 0.000 0.128 0.145
#> x4 -0.006 0.004 -0.006 -1.647 0.100 -0.014 0.001
#> ---------------------------------------------------------------------------------------
#>
# use index of variable instead of name
ols_step_both_p(model, exclude = c(1))
#>
#>
#> Stepwise Summary
#> --------------------------------------------------------------------------------
#> Step Variable AIC SBC SBIC R2 Adj. R2
#> --------------------------------------------------------------------------------
#> 0 Base Model 56988.482 57004.289 228.318 0.00000 0.00000
#> 1 x6 (+) 33473.297 33497.007 -23285.069 0.69145 0.69143
#> 2 x3 (+) 32959.771 32991.385 -23798.829 0.69930 0.69927
#> 3 x2 (+) 32080.424 32119.941 -24678.188 0.71226 0.71222
#> 4 x4 (+) 32079.710 32127.131 -24679.159 0.71230 0.71224
#> --------------------------------------------------------------------------------
#>
#> Final Model Output
#> ------------------
#>
#> Model Summary
#> -----------------------------------------------------------------
#> R 0.844 RMSE 0.539
#> R-Squared 0.712 MSE 0.291
#> Adj. R-Squared 0.712 Coef. Var 6893.511
#> Pred R-Squared 0.712 AIC 32079.710
#> MAE 0.430 SBC 32127.131
#> -----------------------------------------------------------------
#> RMSE: Root Mean Square Error
#> MSE: Mean Square Error
#> MAE: Mean Absolute Error
#> AIC: Akaike Information Criteria
#> SBC: Schwarz Bayesian Criteria
#>
#> ANOVA
#> ---------------------------------------------------------------------------
#> Sum of
#> Squares DF Mean Square F Sig.
#> ---------------------------------------------------------------------------
#> Regression 14408.579 4 3602.145 12376.147 0.0000
#> Residual 5819.654 19995 0.291
#> Total 20228.233 19999
#> ---------------------------------------------------------------------------
#>
#> Parameter Estimates
#> ---------------------------------------------------------------------------------------
#> model Beta Std. Error Std. Beta t Sig lower upper
#> ---------------------------------------------------------------------------------------
#> (Intercept) -0.004 0.004 -1.006 0.315 -0.011 0.004
#> x6 0.148 0.002 0.528 59.728 0.000 0.143 0.153
#> x3 0.141 0.005 0.199 31.259 0.000 0.132 0.150
#> x2 0.137 0.005 0.190 29.983 0.000 0.128 0.145
#> x4 -0.006 0.004 -0.006 -1.647 0.100 -0.014 0.001
#> ---------------------------------------------------------------------------------------
#>
# steps
ols_step_both_p(model, steps = 3)
#>
#>
#> Stepwise Summary
#> --------------------------------------------------------------------------------
#> Step Variable AIC SBC SBIC R2 Adj. R2
#> --------------------------------------------------------------------------------
#> 0 Base Model 56988.482 57004.289 228.318 0.00000 0.00000
#> 1 x6 (+) 33473.297 33497.007 -23285.069 0.69145 0.69143
#> 2 x1 (+) 32931.758 32963.372 -23826.833 0.69972 0.69969
#> 3 x3 (+) 31912.722 31952.239 -24845.827 0.71466 0.71462
#> --------------------------------------------------------------------------------
#>
#> Final Model Output
#> ------------------
#>
#> Model Summary
#> -----------------------------------------------------------------
#> R 0.845 RMSE 0.537
#> R-Squared 0.715 MSE 0.289
#> Adj. R-Squared 0.715 Coef. Var 6864.964
#> Pred R-Squared 0.715 AIC 31912.722
#> MAE 0.428 SBC 31952.239
#> -----------------------------------------------------------------
#> RMSE: Root Mean Square Error
#> MSE: Mean Square Error
#> MAE: Mean Absolute Error
#> AIC: Akaike Information Criteria
#> SBC: Schwarz Bayesian Criteria
#>
#> ANOVA
#> ---------------------------------------------------------------------------
#> Sum of
#> Squares DF Mean Square F Sig.
#> ---------------------------------------------------------------------------
#> Regression 14456.391 3 4818.797 16694.265 0.0000
#> Residual 5771.842 19996 0.289
#> Total 20228.233 19999
#> ---------------------------------------------------------------------------
#>
#> Parameter Estimates
#> ---------------------------------------------------------------------------------------
#> model Beta Std. Error Std. Beta t Sig lower upper
#> ---------------------------------------------------------------------------------------
#> (Intercept) -0.005 0.004 -1.224 0.221 -0.012 0.003
#> x6 0.142 0.002 0.507 57.014 0.000 0.137 0.147
#> x1 0.148 0.004 0.210 32.815 0.000 0.139 0.156
#> x3 0.146 0.005 0.205 32.363 0.000 0.137 0.155
#> ---------------------------------------------------------------------------------------
#>
# }