DMR: Delete or Merge Regressors

In DMR: Delete or Merge Regressors for linear model selection.

Description

DMR is a stepwise backward model selection procedure which simultaneously deletes continuous variables and merges levels of factors. It is based on ranking linear hypotheses with squared t-statistics, using hierarchical clustering for each categorical variable. The final model is selected by minimization of generalized information criterion in the nested family of models.

Usage

 DMR(model, K = log(nrow(model$model)), clust.method = "complete") 

Arguments

model	initial model of class lm.
K	penalty for the number of parameters in generalized information criterion, default is log(n).
clust.method	method of clustering the same as in hclust.

Value

a list including elements

Partitions	a list of partitions of factors for the models on the nested path searched through
Crit	values of generalized information criterion for the models on the nested path searched through
LogLik	values of log-likelihood for the models on the nested path searched through
Best	a list containing features of the selected model: Partition, Model of class lm, Crit and Hypotheses represesnted as a matrix of lienear hypotheses imposed on the model's parameters

Author(s)

Piotr Pokarowski, Agnieszka Prochenka, Aleksandra Maj

See Also

stepDMR, DMR4glm, plot_bf, roc

Examples

k=4
v1 <- factor(rep(1:8, each = 12*k))
v2 <- factor(rep(1:4, times = 24*k))
v3 <- factor(rep(1:3, times = 32*k))
x1 <- rnorm(96*k)
x2 <- runif(96*k)
y <- rep(c(2, 2, -1, -1, -1, -1, 0, 0), each = 12*k) + rnorm(96*k)
m <- lm(y ~ v1 + v2 + v3 + x1 + x2)
(out <- DMR(m))

Example output

Loading required package: magic
Loading required package: abind
$Partitions
$Partitions[[1]]
$Partitions[[1]]$v1
1 2 3 4 5 6 7 8 
1 2 3 4 5 6 7 8 

$Partitions[[1]]$v2
1 2 3 4 
1 2 3 4 

$Partitions[[1]]$v3
1 2 3 
1 2 3 


$Partitions[[2]]
$Partitions[[2]]$v1
1 2 3 4 5 6 7 8 
1 2 3 4 5 6 7 8 

$Partitions[[2]]$v2
1 2 3 4 
1 2 3 2 

$Partitions[[2]]$v3
1 2 3 
1 2 3 


$Partitions[[3]]
$Partitions[[3]]$v1
1 2 3 4 5 6 7 8 
1 2 3 4 5 6 7 8 

$Partitions[[3]]$v2
1 2 3 4 
1 2 2 2 

$Partitions[[3]]$v3
1 2 3 
1 2 3 


$Partitions[[4]]
$Partitions[[4]]$v1
1 2 3 4 5 6 7 8 
1 2 3 4 5 5 6 7 

$Partitions[[4]]$v2
1 2 3 4 
1 2 2 2 

$Partitions[[4]]$v3
1 2 3 
1 2 3 


$Partitions[[5]]
$Partitions[[5]]$v1
1 2 3 4 5 6 7 8 
1 2 3 4 5 5 6 7 

$Partitions[[5]]$v2
1 2 3 4 
1 2 2 2 

$Partitions[[5]]$v3
1 2 3 
1 2 3 


$Partitions[[6]]
$Partitions[[6]]$v1
1 2 3 4 5 6 7 8 
1 1 2 3 4 4 5 6 

$Partitions[[6]]$v2
1 2 3 4 
1 2 2 2 

$Partitions[[6]]$v3
1 2 3 
1 2 3 


$Partitions[[7]]
$Partitions[[7]]$v1
1 2 3 4 5 6 7 8 
1 1 2 3 3 3 4 5 

$Partitions[[7]]$v2
1 2 3 4 
1 2 2 2 

$Partitions[[7]]$v3
1 2 3 
1 2 3 


$Partitions[[8]]
$Partitions[[8]]$v1
1 2 3 4 5 6 7 8 
1 1 2 3 3 3 4 5 

$Partitions[[8]]$v2
1 2 3 4 
1 2 2 2 

$Partitions[[8]]$v3
1 2 3 
1 2 2 


$Partitions[[9]]
$Partitions[[9]]$v1
1 2 3 4 5 6 7 8 
1 1 2 3 3 3 4 4 

$Partitions[[9]]$v2
1 2 3 4 
1 2 2 2 

$Partitions[[9]]$v3
1 2 3 
1 2 2 


$Partitions[[10]]
$Partitions[[10]]$v1
1 2 3 4 5 6 7 8 
1 1 2 3 3 3 4 4 

$Partitions[[10]]$v2
1 2 3 4 
1 2 2 2 

$Partitions[[10]]$v3
1 2 3 
1 2 2 


$Partitions[[11]]
$Partitions[[11]]$v1
1 2 3 4 5 6 7 8 
1 1 2 3 3 3 4 4 

$Partitions[[11]]$v2
1 2 3 4 
1 1 1 1 

$Partitions[[11]]$v3
1 2 3 
1 2 2 


$Partitions[[12]]
$Partitions[[12]]$v1
1 2 3 4 5 6 7 8 
1 1 2 3 3 3 4 4 

$Partitions[[12]]$v2
1 2 3 4 
1 1 1 1 

$Partitions[[12]]$v3
1 2 3 
1 1 1 


$Partitions[[13]]
$Partitions[[13]]$v1
1 2 3 4 5 6 7 8 
1 1 2 2 2 2 3 3 

$Partitions[[13]]$v2
1 2 3 4 
1 1 1 1 

$Partitions[[13]]$v3
1 2 3 
1 1 1 


$Partitions[[14]]
$Partitions[[14]]$v1
1 2 3 4 5 6 7 8 
1 1 2 2 2 2 2 2 

$Partitions[[14]]$v2
1 2 3 4 
1 1 1 1 

$Partitions[[14]]$v3
1 2 3 
1 1 1 


$Partitions[[15]]
$Partitions[[15]]$v1
[1] 1 1 1 1 1 1 1 1

$Partitions[[15]]$v2
[1] 1 1 1 1

$Partitions[[15]]$v3
[1] 1 1 1



$Crit
 [1] 1121.893 1115.945 1110.039 1104.151 1098.338 1092.576 1086.860 1081.441
 [9] 1076.060 1072.057 1068.429 1064.365 1060.997 1139.985 1421.514

$LogLik
 [1] -513.3413 -513.3427 -513.3650 -513.3961 -513.4653 -513.5592 -513.6770
 [8] -513.9428 -514.2274 -515.2013 -516.3626 -517.3057 -518.5974 -561.0665
[15] -704.8064

$Best
$Best$Partition
$Best$Partition$v1
1 2 3 4 5 6 7 8 
1 1 2 2 2 2 3 3 

$Best$Partition$v2
1 2 3 4 
1 1 1 1 

$Best$Partition$v3
1 2 3 
1 1 1 


$Best$Model

Call:
lm(formula = y ~ ., data = newdata)

Coefficients:
(Intercept)          v12          v13  
      1.909       -2.922       -1.784  


$Best$Crit
[1] 1060.997

$Best$Hypotheses
      Intercept v1 v1 v1 v1 v1 v1 v1 v2 v2 v2 v3 v3 x1 x2
 [1,]         0  0  0  0  0  0  0  0 -1  0  1  0  0  0  0
 [2,]         0  0  0  0  0  0  0  0 -1  1  0  0  0  0  0
 [3,]         0  0  0  0 -1  1  0  0  0  0  0  0  0  0  0
 [4,]         0  0  0  0  0  0  0  0  0  0  0  0  0  0  1
 [5,]         0  1  0  0  0  0  0  0  0  0  0  0  0  0  0
 [6,]         0  0  0 -1  1  0  0  0  0  0  0  0  0  0  0
 [7,]         0  0  0  0  0  0  0  0  0  0  0 -1  1  0  0
 [8,]         0  0  0  0  0  0 -1  1  0  0  0  0  0  0  0
 [9,]         0  0  0  0  0  0  0  0  0  0  0  0  0  1  0
[10,]         0  0  0  0  0  0  0  0  1  0  0  0  0  0  0
[11,]         0  0  0  0  0  0  0  0  0  0  0  1  0  0  0
[12,]         0  0 -1  1  0  0  0  0  0  0  0  0  0  0  0


Warning messages:
1: In sigma_sq * (apply(x, 1, function(y) t(y) %*% y)) :
  Recycling array of length 1 in array-vector arithmetic is deprecated.
  Use c() or as.vector() instead.

2: In sigma_sq * (apply(x, 1, function(y) t(y) %*% y)) :
  Recycling array of length 1 in array-vector arithmetic is deprecated.
  Use c() or as.vector() instead.

DMR documentation built on May 30, 2017, 6:25 a.m.