Voting rights

Step 3 in pipeline

Voting rights are assigned per user and per entity, based on users' trust scores and privacy settings.

VotingRightsAssignment

Bases: ABC

__call__

__call__(users: DataFrame, entities: DataFrame, vouches: DataFrame, privacy: PrivacySettings, user_models: dict[int, ScoringModel]) -> tuple[VotingRights, DataFrame]

Compute voting rights

Parameters:

Name	Type	Description	Default
users	DataFrame	user_id (int, index) trust_score (float)	required
entities	DataFrame	entity_id (int, index)	required
vouches	DataFrame	voucher (int) vouchee (int) vouch (float)	required
privacy	PrivacySettings	privacy[user, entity] in { True, False, None }	required
user_models	dict[int, ScoringModel]	user_models[user_id] is the user's scoring model	required

Returns:

Name	Type	Description
	voting_rights[user, entity] is the voting right	of a user on entity for criterion
entities	DataFrame with columns	entity_id (int, index)

to_json

to_json()

IsTrust

IsTrust(privacy_penalty: float = 0.5)

Bases: VotingRightsAssignment

Implements voting rights assignment directly based on user trust scores and applying some privacy penalty.

Parameters:

Name	Type	Description	Default
privacy_penalty	float	Penalty on private comparisons.	0.5

privacy_penalty

privacy_penalty = privacy_penalty

__call__

__call__(users: DataFrame, entities: DataFrame, vouches: DataFrame, privacy: PrivacySettings, user_models: dict[int, ScoringModel]) -> tuple[VotingRights, DataFrame]

Compute voting rights

Parameters:

Name	Type	Description	Default
users	DataFrame	user_id (int, index) trust_score (float)	required
entities	DataFrame	entity_id (int, index)	required
vouches	DataFrame		required
privacy	PrivacySettings	privacy[user, entity] in { True, False, None }	required
user_models	dict[int, ScoringModel]		required

Returns:

Name	Type	Description
	voting_rights[user, entity] is the voting right	of a user on entity for criterion
entities	DataFrame with columns	entity_id (int, index)

to_json

to_json()

AffineOvertrust

AffineOvertrust(privacy_penalty: float = 0.5, min_overtrust: float = 2.0, overtrust_ratio: float = 0.1)

Bases: VotingRightsAssignment

Initializes the AffineOvertrust model with the specified parameters.

Parameters:

Name	Type	Description	Default
privacy_penalty	float	Penalty on private comparisons.	0.5
min_overtrust	float	Minimum overtrust amount to distribute over all users.	2.0
overtrust_ratio	float	Trust ratio used to compute maximal allowed overtrust.	0.1

privacy_penalty

privacy_penalty = privacy_penalty

min_overtrust

min_overtrust = min_overtrust

overtrust_ratio

overtrust_ratio = overtrust_ratio

__call__

__call__(users: DataFrame, entities: DataFrame, vouches: DataFrame, privacy: PrivacySettings, user_models: Optional[dict[int, ScoringModel]]) -> tuple[VotingRights, DataFrame]

Compute voting rights for users on entities

Parameters:

Name	Type	Description	Default
users	DataFrame	user_id (int, index) trust_score (float)	required
entities	DataFrame	entity_id (int, index)	required
vouches	DataFrame		required
privacy	PrivacySettings	privacy[user, entity] is the privacy setting of user for entity May be True, False or None	required
user_models	Optional[dict[int, ScoringModel]]		required

Returns:

Name	Type	Description
	voting_rights[user, entity] is the voting right	of a user on entity for criterion
entities	DataFrame with columns	entity_id (int, index) cumulative_trust (float) min_voting_right (float) overtrust (float)

compute_entity_voting_rights

compute_entity_voting_rights(trust_scores: Series, privacy_weights: Series) -> tuple[Series, float, float, float]

cumulative_trust

cumulative_trust(trust_scores: ndarray, privacy_weights: ndarray) -> float

Returns the sum of trusts of raters of entity entity_id, weighted by their privacy setting.

Parameters:

Name	Type	Description	Default
trust_scores	ndarray	index "user_id (int, index) values "trust_score" (float)	required
privacy_weights	ndarray	privacy_weights[u] is the privacy weight of user u	required

Returns:

Name	Type	Description
out	float

maximal_overtrust

maximal_overtrust(trust: float) -> float

Computes the maximal allowed overtrust of an entity, for a given total trust of the entity's raters.

Parameters:

Name	Type	Description	Default
trust	float	Cumulative trust received by the entity	required

Returns:

Name	Type	Description
out	float

overtrust

overtrust(min_voting_right: float, trust_scores: ndarray, privacy_weights: ndarray) -> float

Returns the overtrust, if min_voting_right is enforced upon all raters.

Parameters:

Name	Type	Description	Default
min_voting_right	float	Overtrust for min_voting_right	required
trust_scores	ndarray	values: trust_score (float)	required
privacy_weights	ndarray	values: privacy weight of user	required

Returns:

Name	Type	Description
out	float

min_voting_right

min_voting_right(max_overtrust: float, trust_scores: ndarray, privacy_weights: ndarray) -> float

Returns the minimal voting rights that corresponds to max_overtrust.

Parameters:

Name	Type	Description	Default
max_overtrust	float	Maximal overtrust allowed for entity_id	required
trust_scores	ndarray	trust score values per user	required
privacy_weights	ndarray	privacy weight per user	required

Returns:

Name	Type	Description
out	float

to_json

to_json()