`colley_ranker`

Colley pairwise preference ranker.

The library implements the ranker_protocol defined in the ranking_protocols library. It provides predicates for learning a ranker from pairwise preferences, using it to order candidate items, and exporting it as a list of predicate clauses or to a file.

Datasets are represented as objects implementing the pairwise_ranking_dataset_protocol protocol from the ranking_protocols library. See the test_datasets directory for examples. Requires well-formed connected pairwise datasets so that the learned rankings remain globally comparable across all ranked items.

API documentation

Open the ../../apis/library_index.html#colley_ranker link in a web browser.

Loading

To load this library, load the loader.lgt file:

| ?- logtalk_load(colley_ranker(loader)).

Testing

To test this library predicates, load the tester.lgt file:

| ?- logtalk_load(colley_ranker(tester)).

Features

Pairwise Preference Learning: Learns one deterministic rating per item from aggregated pairwise outcomes.
Colley Matrix Fidelity: Solves the standard Colley linear system with diagonal regularization 2 + games_i and off-diagonal entries -games_ij.
Numerically Hardened Solver: Uses Gaussian elimination with partial pivoting plus residual checks before accepting the learned ratings.
Conservative Ratings: Produces ratings in the interval [0,1] with the neutral prior centered at 0.5.
Deterministic Ranking: Orders candidate items by learned rating with deterministic tie-breaking.
Strict Dataset Validation: Rejects duplicate items, undeclared items, self-preferences, non-positive weights, and disconnected comparison graphs.
Ranker Export: Learned rankers can be exported as self-contained terms.
Shared Ranking Infrastructure: Reuses the ranking_protocols helpers for dataset validation, diagnostics, export, and candidate ranking.

Scoring semantics

This implementation aggregates pairwise preferences into matchup totals and then solves the Colley system

C r = b

where C_ii = 2 + games_i, C_ij = -games_ij for i \= j, and b_i = 1 + (wins_i - losses_i) / 2.

The linear system is solved using deterministic Gaussian elimination with partial pivoting. The implementation also verifies that the recovered solution has a small residual and only clamps negligible floating-point boundary noise at 0.0 or 1.0; materially invalid out-of-range values are rejected instead of being silently accepted.

The resulting ratings remain close to 0.5 when evidence is weak and move toward 1.0 or 0.0 only when repeated pairwise results justify doing so.

Usage

Learning a ranker

% Learn from a pairwise ranking dataset object
| ?- colley_ranker::learn(my_dataset, Ranker).
...

% Learn with an explicit empty options list
| ?- colley_ranker::learn(my_dataset, Ranker, []).
...

The current implementation accepts only the empty options list []. Any non-empty options list is rejected.

Inspecting diagnostics

% Inspect model and dataset summary metadata
| ?- colley_ranker::learn(my_dataset, Ranker),
     colley_ranker::diagnostics(Ranker, Diagnostics).
Diagnostics = [...]
...

Ranking candidate items

% Rank a candidate set from most preferred to least preferred
| ?- colley_ranker::learn(my_dataset, Ranker),
     colley_ranker::rank(Ranker, [item_a, item_b, item_c], Ranking).
Ranking = [...]
...

Candidate lists must be proper lists of unique, ground items declared by the training dataset. Invalid ranker terms, duplicate candidates, and candidates containing variables are rejected with errors instead of being silently accepted.

Exporting the ranker

Learned rankers can be exported as a list of clauses or to a file for later use.

% Export as predicate clauses
| ?- colley_ranker::learn(my_dataset, Ranker),
     colley_ranker::export_to_clauses(my_dataset, Ranker, my_ranker, Clauses).
Clauses = [my_ranker(colley_ranker(...))]
...

% Export to a file
| ?- colley_ranker::learn(my_dataset, Ranker),
     colley_ranker::export_to_file(my_dataset, Ranker, my_ranker, 'ranker.pl').
...

Diagnostics syntax

The diagnostics/2 predicate returns a list of metadata terms with the form:

[
    model(colley_ranker),
    options(Options),
    dataset_summary(DatasetSummary)
]

Ranker representation

The learned ranker is represented by a compound term of the form:

colley_ranker(Items, Ratings, Diagnostics)

Where:

Items: List of ranked items.
Scores: List of Item-Rating pairs.
Diagnostics: List of metadata terms, including the effective options and dataset summary.

References

Colley, W. N. (2002). Colley’s bias free college football ranking method: The Colley Matrix explained.

colley_ranker