Skip to contents

The same I and O operations as Hits, but the loop stops when the largest absolute change in the normalised hub vector falls below tol, which is the "converge" wording of Kleinberg's Iterate procedure rather than a fixed sweep count.

Usage

Hubsa(y, A, tol = 1e-12, max_iter = 1000)

Arguments

y

Starting hub vector of length n; ones reproduce the paper.

A

Square adjacency matrix.

tol

Positive convergence tolerance on the hub vector.

max_iter

Iteration cap.

Value

List with estimate, hubs, authorities, iterations, delta, converged, n, method.

Details

Formula: repeat a <- A' h, h <- A a until max|dh| <= tol.

References

Kleinberg (1999), Journal of the ACM 46(5):604-632. doi:10.1145/324133.324140

Examples

Hubsa(y = c(1, 1, 1), A = matrix(c(0, 1, 1, 0, 0, 1, 1, 0, 0), 3, 3, byrow = TRUE))
#> $estimate
#> [1] 0.8506508
#> 
#> $hubs
#> [1] 8.506508e-01 5.257311e-01 5.495129e-13
#> 
#> $authorities
#> [1] 8.891306e-13 5.257311e-01 8.506508e-01
#> 
#> $iterations
#> [1] 29
#> 
#> $delta
#> [1] 8.891306e-13
#> 
#> $converged
#> [1] TRUE
#> 
#> $n
#> [1] 3
#> 
#> $method
#> [1] "Kleinberg (1999) I/O recursion, stopped at max|dh| <= tol"
#> 
#> attr(,"class")
#> [1] "morie_rich_result" "list"