design-matrix.R

# create dummy data -------------------------------------------------------
set.seed(1)
d <- data.frame(col1 = sample(letters[1:3], 10, replace=T), 
                col2 = sample(letters[24:26], 10, replace=T),
                col3 = runif(10) * 10, 
                stringsAsFactors = FALSE)

d$col1 <- as.factor(d$col1)
d$col4 = d$col3 + runif(10)

d; str(d)
#    col1 col2      col3      col4
# 1     a    x 9.3470523 9.8291324
# 2     b    x 2.1214252 2.7209910
# 3     b    z 6.5167377 7.0102790
# 4     c    y 1.2555510 1.4417686
# 5     a    z 2.6722067 3.4995800
# 6     c    y 3.8611409 4.5296077
# 7     c    z 0.1339033 0.9281432
# 8     b    z 3.8238796 3.9318232
# 9     b    y 8.6969085 9.4206194
# 10    a    z 3.4034900 3.8147644
# 'data.frame':	10 obs. of  4 variables:
#   $ col1: Factor w/ 3 levels "a","b","c": 1 2 2 3 1 3 3 2 2 1
#   $ col2: chr  "x" "x" "z" "y" ...
#   $ col3: num  9.35 2.12 6.52 1.26 2.67 ...
#   $ col4: num  9.83 2.72 7.01 1.44 3.5 ...


# data.matrix -------------------------------------------------------------
# data.matrix converts dataframe into numerical matrix
data.matrix(d)  # string columns return NA
#       col1 col2      col3      col4
#  [1,]    1   NA 9.3470523 9.8291324
#  [2,]    2   NA 2.1214252 2.7209910
#  [3,]    2   NA 6.5167377 7.0102790
#  [4,]    3   NA 1.2555510 1.4417686
#  [5,]    1   NA 2.6722067 3.4995800
#  [6,]    3   NA 3.8611409 4.5296077
#  [7,]    3   NA 0.1339033 0.9281432
#  [8,]    2   NA 3.8238796 3.9318232
#  [9,]    2   NA 8.6969085 9.4206194
# [10,]    1   NA 3.4034900 3.8147644


# model.matrix ------------------------------------------------------------

model.matrix(~col3, data=d)
#    (Intercept)      col3
# 1            1 9.3470523
# 2            1 2.1214252
# 3            1 6.5167377
# 4            1 1.2555510
# 5            1 2.6722067
# 6            1 3.8611409
# 7            1 0.1339033
# 8            1 3.8238796
# 9            1 8.6969085
# 10           1 3.4034900
# attr(,"assign")
# [1] 0 1

model.matrix(col4~col3, data=d)  # left side is ignored
#    (Intercept)      col3
# 1            1 9.3470523
# 2            1 2.1214252
# 3            1 6.5167377
# 4            1 1.2555510
# 5            1 2.6722067
# 6            1 3.8611409
# 7            1 0.1339033
# 8            1 3.8238796
# 9            1 8.6969085
# 10           1 3.4034900
# attr(,"assign")
# [1] 0 1

model.matrix(~col1+col2, data=d)  # intercept replaces 1 column (col1a)
#    (Intercept) col1b col1c col2y col2z
# 1            1     0     0     0     0
# 2            1     1     0     0     0
# 3            1     1     0     0     1
# 4            1     0     1     1     0
# 5            1     0     0     0     1
# 6            1     0     1     1     0
# 7            1     0     1     0     1
# 8            1     1     0     0     1
# 9            1     1     0     1     0
# 10           1     0     0     0     1
# attr(,"assign")
# [1] 0 1 1 2 2
# attr(,"contrasts")
# attr(,"contrasts")$col1
# [1] "contr.treatment"
# 
# attr(,"contrasts")$col2
# [1] "contr.treatment"

model.matrix(~col1+col2-1, data=d)  # "+0" or "-1" drops intercept
#    col1a col1b col1c col2y col2z
# 1      1     0     0     0     0
# 2      0     1     0     0     0
# 3      0     1     0     0     1
# 4      0     0     1     1     0
# 5      1     0     0     0     1
# 6      0     0     1     1     0
# 7      0     0     1     0     1
# 8      0     1     0     0     1
# 9      0     1     0     1     0
# 10     1     0     0     0     1
# attr(,"assign")
# [1] 1 1 1 2 2
# attr(,"contrasts")
# attr(,"contrasts")$col1
# [1] "contr.treatment"
# 
# attr(,"contrasts")$col2
# [1] "contr.treatment"

model.matrix(col4~.+0-col3, data=d)  # same results as above
#    col1a col1b col1c col2y col2z
# 1      1     0     0     0     0
# 2      0     1     0     0     0
# 3      0     1     0     0     1
# 4      0     0     1     1     0
# 5      1     0     0     0     1
# 6      0     0     1     1     0
# 7      0     0     1     0     1
# 8      0     1     0     0     1
# 9      0     1     0     1     0
# 10     1     0     0     0     1
# attr(,"assign")
# [1] 1 1 1 2 2
# attr(,"contrasts")
# attr(,"contrasts")$col1
# [1] "contr.treatment"
# 
# attr(,"contrasts")$col2
# [1] "contr.treatment"


# one hot encoding --------------------------------------------------------

# create new contrast function with 'contrasts=FALSE'
contr.onehot <- function (n, contrasts, sparse=FALSE) {
  contr.sum(n=n, contrasts=FALSE, sparse=sparse)
}

# set options
options(contrasts = c("contr.onehot", "contr.onehot"))

# use model.matrix function
model.matrix(~col1+col2-1, data=d)
#    col1a col1b col1c col2x col2y col2z
# 1      1     0     0     1     0     0
# 2      0     1     0     1     0     0
# 3      0     1     0     0     0     1
# 4      0     0     1     0     1     0
# 5      1     0     0     0     0     1
# 6      0     0     1     0     1     0
# 7      0     0     1     0     0     1
# 8      0     1     0     0     0     1
# 9      0     1     0     0     1     0
# 10     1     0     0     0     0     1
# attr(,"assign")
# [1] 1 1 1 2 2 2
# attr(,"contrasts")
# attr(,"contrasts")$col1
# [1] "contr.onehot"
# 
# attr(,"contrasts")$col2
# [1] "contr.onehot"