Ctrl+K

Unanimous Models

Unanimous Models#

Identity samplers are not fascinating per se as all voters have the same preferences. There are useful tools however, for instance when using them in mixtures.

identity(num_voters: int, num_candidates: int, rel_num_approvals: float, seed: int = None) → list[set[int]][source]#

Generates approval votes from the identity culture. These votes are simples: all voters approves of the candidates 0, 1, 2, …, ⌊rel_num_approvals * num_candidates⌋ and only these ones.

Parameters:

  • num_voters (int) – Number of Voters.

  • num_candidates (int) – Number of Candidates.

  • rel_num_approvals (float) – Proportion of approved candidates in a ballot.

  • seed (int) – Seed for numpy random number generator.

Returns:

Approval votes.

Return type:

list[set[int]]

Examples

from prefsampling.approval import identity

# Sample a unanimous profile with 2 voters and 3 candidates.
# Voters approve 60% of the candidates (1 in this case).
identity(2, 3, 0.6)

# The seed will not change anything here, but you can still set it.
identity(2, 3, 0.6, seed=1002)

# Parameter rel_num_approvals needs to be in [0, 1]
try:
    identity(2, 3, 1.6)
except ValueError:
    pass
try:
    identity(2, 3, -0.6)
except ValueError:
    pass

Validation

Validation is trivial here, we thus omit it.

empty(num_voters: int, num_candidates: int, seed: int = None) → list[set[int]][source]#

Generates approval votes where each vote is empty.

Parameters:

  • num_voters (int) – Number of Voters.

  • num_candidates (int) – Number of Candidates.

  • seed (int) – Seed for numpy random number generator.

Returns:

Approval votes.

Return type:

list[set[int]]

Examples

from prefsampling.approval import empty

# "Sample" a profile of 2 voters approving of no candidates.
# The number of candidates is not used but must be given to keep
# the samplers' signatures consistent.
empty(2, 3)

# The seed will not change anything here, but you can still set it.
empty(2, 3, seed=1002)

Validation

Validation is trivial here, we thus omit it.

full(num_voters: int, num_candidates: int, seed: int = None) → list[set[int]][source]#

Generates approval votes where all voters approve of all the candidates.

Parameters:

  • num_voters (int) – Number of Voters.

  • num_candidates (int) – Number of Candidates.

  • seed (int) – Seed for numpy random number generator.

Returns:

Approval votes.

Return type:

list[set[int]]

Examples

from prefsampling.approval import full

# "Sample" a profile of 2 voters approving all candidates.
full(2, 3)

# The seed will not change anything here, but you can still set it.
full(2, 3, seed=1002)

Validation

Validation is trivial here, we thus omit it.

Contents