Covariance matrix for preferences, calculated using the rankings matrix.

Usage

pref_cov(x, preferences_col = NULL, frequency_col = NULL, ...)

Arguments

x: A vector of preferences, or a tibble with a column of preferences.
preferences_col: <tidy-select> When x is a tibble, the column containing the preferences.
frequency_col: <tidy-select> When x is a tibble, the column containing the frequency of the preferences. If not provided, each row is considered to be observed a single time.
...: Extra arguments to be passed to stats::cov.wt.

Value

A covariance matrix containing covariances for the ranks assigned to item pairs.

Examples

# Simple covariance on a vector of preferences
prefs <- preferences(c("a > b > c", "b > c > a", "c > a > b"))
pref_cov(prefs)
#> $cov
#>      a    b    c
#> a  1.0 -0.5 -0.5
#> b -0.5  1.0 -0.5
#> c -0.5 -0.5  1.0
#> 
#> $center
#> a b c 
#> 2 2 2 
#> 
#> $n.obs
#> [1] 3
#> 

# Weighted covariance by frequency
df <- tibble::tibble(
  prefs = preferences(c("a > b > c", "b > c > a")),
  freq = c(3, 2)
)
pref_cov(df, preferences_col = prefs, frequency_col = freq)
#> $cov
#>      a    b    c
#> a  1.2 -0.6 -0.6
#> b -0.6  0.3  0.3
#> c -0.6  0.3  0.3
#> 
#> $center
#>   a   b   c 
#> 1.8 1.6 2.6 
#> 
#> $n.obs
#> [1] 5
#>