Csiszar (1967) f-divergence D_f(p || q) = sum q f(p/q)
Arguments
- p, q
non-negative weights of equal length
- f
a convex generator with f(1) = 0, or one of "kl", "rkl", "tv", "chi2", "hellinger", "js"
- f_inf
recession constant charged against mass p places where q is zero; defaults to the exact value for a named generator
- normalise
divide p and q by their totals first
Examples
V <- c(1, 2, 3, 4, 5, 6, 7, 8)
FDiverg(V, V)
#> $divergence
#> [1] 0
#>
#> $estimate
#> [1] 0
#>
#> $terms
#> [1] 0 0 0 0 0 0 0 0
#>
#> $support
#> [1] 0
#>
#> $generator
#> [1] "kl"
#>
#> $f_inf
#> [1] Inf
#>
#> $n
#> [1] 8
#>
#> $method
#> [1] "Csiszar (1967) f-divergence, sum q f(p/q)"
#>
