Struct RankAggregation 

pub struct RankAggregation;

Rank aggregation methods producing a single consensus ranking.

Implementations

impl RankAggregation

pub fn borda_count(rankings: &[Vec<usize>]) -> SklResult<Vec<usize>>

Borda count rank aggregation.

Each ranking awards n - rank points to the item at a given position (the best item, at rank 0, receives n points; the worst receives 1). Points are summed across all rankings and the consensus ranking lists items in order of decreasing total score. Ties are broken by smaller item index so the output is deterministic.

Every ranking must be a permutation of {0, ..., n-1} for the same n.

pub fn kemeny_optimal(rankings: &[Vec<usize>]) -> SklResult<Vec<usize>>

Kemeny-optimal rank aggregation.

Returns the consensus ranking that minimises the total Kendall-tau distance (number of pairwise disagreements) to the input rankings. For n <= 8 the optimum is found exactly by enumerating all n! permutations. For larger n an exhaustive search is intractable, so a documented local-search heuristic is used instead: starting from the Borda consensus, adjacent items are repeatedly swapped whenever the swap strictly reduces the total Kendall-tau cost, until no improving adjacent swap remains (a local Kemeny optimum). The result is therefore exact for small n and an approximation for large n.

Trait Implementations

impl Clone for RankAggregation

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for RankAggregation

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.