Bradley-Terry¶

evalica.bradley_terry(xs, ys, winners, index=None, weights=None, win_weight=1.0, tie_weight=0.5, solver='pyo3', tolerance=1e-06, limit=100, **kwargs)[source]¶

Compute the Bradley-Terry scores for the given pairwise comparison.

Note

Bradley, R.A., Terry, M.E.: Rank Analysis of Incomplete Block Designs: I. The Method of Paired Comparisons. Biometrika. 39, 324–345 (1952). <https://doi.org/10.2307/2334029>.

Note

Newman, M.E.J.: Efficient Computation of Rankings from Pairwise Comparisons. Journal of Machine Learning Research. 24, 1–25 (2023). <https://www.jmlr.org/papers/v24/22-1086.html>.

Parameters:

xs (Collection[T_contra]) – The left-hand side elements.
ys (Collection[T_contra]) – The right-hand side elements.
winners (Collection[Winner]) – The winner elements.
index (Index | None) – The index.
weights (Collection[float] | None) – The example weights.
win_weight (float) – The win weight.
tie_weight (float) – The tie weight.
solver (Literal['naive', 'pyo3']) – The solver.
tolerance (float) – The convergence tolerance.
limit (int) – The maximum number of iterations.
**kwargs (Any) – The additional arguments.

Returns:

The Bradley-Terry result.

Return type:

BradleyTerryResult

class evalica.BradleyTerryResult(scores, index, win_weight, tie_weight, solver, tolerance, iterations, limit)[source]¶

The Bradley-Terry result.

Parameters:

scores (Series)
index (Index)
win_weight (float)
tie_weight (float)
solver (Literal['naive', 'pyo3'])
tolerance (float)
iterations (int)
limit (int)

scores¶

The element scores.

Type:: pandas.Series

index¶

The index.

Type:: pandas.Index

win_weight¶

The win weight.

Type:: float

tie_weight¶

The tie weight.

Type:: float

solver¶

The solver.

Type:: Literal[‘naive’, ‘pyo3’]

tolerance¶

The convergence tolerance.

Type:: float

iterations¶

The actual number of iterations.

Type:: int

limit¶

The maximum number of iterations.

Type:: int

evalica.newman(xs, ys, winners, index=None, v_init=0.5, weights=None, win_weight=1.0, tie_weight=1.0, solver='pyo3', tolerance=1e-06, limit=100, **kwargs)[source]¶

Compute the scores for the given pairwise comparison using the Newman’s algorithm.

Note

Newman, M.E.J.: Efficient Computation of Rankings from Pairwise Comparisons. Journal of Machine Learning Research. 24, 1–25 (2023). <https://www.jmlr.org/papers/v24/22-1086.html>.

Parameters:

xs (Collection[T_contra]) – The left-hand side elements.
ys (Collection[T_contra]) – The right-hand side elements.
winners (Collection[Winner]) – The winner elements.
index (Index | None) – The index.
v_init (float) – The initial tie parameter.
weights (Collection[float] | None) – The example weights.
win_weight (float) – The win weight.
tie_weight (float) – The tie weight.
solver (Literal['naive', 'pyo3']) – The solver.
tolerance (float) – The convergence tolerance.
limit (int) – The maximum number of iterations.
**kwargs (Any) – The additional arguments.

Returns:

The Newman’s result.

Return type:

NewmanResult

class evalica.NewmanResult(scores, index, v, v_init, win_weight, tie_weight, solver, tolerance, iterations, limit)[source]¶

The Newman’s algorithm result.

Parameters:

scores (Series)
index (Index)
v (float)
v_init (float)
win_weight (float)
tie_weight (float)
solver (Literal['naive', 'pyo3'])
tolerance (float)
iterations (int)
limit (int)

scores¶

The element scores.

Type:: pandas.Series

index¶

The index.

Type:: pandas.Index

v¶

The tie parameter.

Type:: float

v_init¶

The initial tie parameter.

Type:: float

win_weight¶

The win weight.

Type:: float

tie_weight¶

The tie weight.

Type:: float

solver¶

The solver.

Type:: Literal[‘naive’, ‘pyo3’]

tolerance¶

The convergence tolerance.

Type:: float

iterations¶

The actual number of iterations.

Type:: int

limit¶

The maximum number of iterations.

Type:: int