nsdlib.common.models module

class nsdlib.common.models.ClassificationMetrics(TP: int, TN: int, FP: int, FN: int, P: int, N: int)[source]

Bases: object

Confusion matrix representation.

It is based on https://en.wikipedia.org/wiki/Confusion_matrix.

property ACC: Accuracy (ACC).

property F1: F1 score.

property FDR: False discovery rate (FDR).

FN: int

property FNR: Miss rate or false negative rate (FNR).

property FOR: False omission rate (FOR).

FP: int

property FPR: Fall-out or false positive rate (FPR).

N: int

property NPV: Negative predictive value (NPV).

P: int

property PPV: Precision or positive predictive value (PPV).

TN: int

property TNR: Specificity, selectivity or true negative rate (TNR).

TP: int

property TPR: Sensitivity, recall, hit rate, or true positive rate (TPR).

property TS: False omission rate (FOR).

property confusion_matrix: List[List[float]]: Confusion matrix.

get_classification_report() → Dict[str, float][source]: Classification report as string.

class nsdlib.common.models.EnsembleSourceDetectionConfig(detection_configs: ~typing.List[~nsdlib.common.models.SourceDetectionConfig] = <factory>, voting_type: ~nsdlib.taxonomies.EnsembleVotingType = EnsembleVotingType.HARD, classifier_weights: ~typing.List[float] = <factory>)[source]

Bases: object

Ensemble source detection configuration.

classifier_weights: List[float]

detection_configs: List[SourceDetectionConfig]

voting_type: EnsembleVotingType = 'hard'

class nsdlib.common.models.EnsembleSourceDetectionResult(config: nsdlib.common.models.EnsembleSourceDetectionConfig, G: networkx.classes.graph.Graph, IG: networkx.classes.graph.Graph, global_scores: Dict[int | str, float], ensemble_scores: List[nsdlib.common.models.SourceDetectionResult], detected_sources: List[str | int])[source]

Bases: object

G: Graph

IG: Graph

config: EnsembleSourceDetectionConfig

detected_sources: List[str | int]

ensemble_scores: List[SourceDetectionResult]

global_scores: Dict[int | str, float]

class nsdlib.common.models.SelectionAlgorithm(selection_method: nsdlib.taxonomies.NodeEvaluationAlgorithm | None = None, selection_threshold: float | None = None)[source]

Bases: object

selection_method: NodeEvaluationAlgorithm | None = None

selection_threshold: float | None = None

class nsdlib.common.models.SourceDetectionConfig(node_evaluation_algorithm: NodeEvaluationAlgorithm = NodeEvaluationAlgorithm.CENTRALITY_DEGREE, selection_algorithm: SelectionAlgorithm | None = None, outbreaks_detection_algorithm: OutbreaksDetectionAlgorithm | None = None, propagation_reconstruction_algorithm: PropagationReconstructionAlgorithm | None = None)[source]

Bases: object

Source detection configuration.

node_evaluation_algorithm: NodeEvaluationAlgorithm = 'degree_centrality'

outbreaks_detection_algorithm: OutbreaksDetectionAlgorithm | None = None

propagation_reconstruction_algorithm: PropagationReconstructionAlgorithm | None = None

selection_algorithm: SelectionAlgorithm | None = None

class nsdlib.common.models.SourceDetectionEvaluation(TP: int, TN: int, FP: int, FN: int, P: int, N: int, real_sources: List[str | int], detected_sources: List[str | int], error_distances: Dict[int | str, float])[source]

Bases: ClassificationMetrics

property avg_error_distance: float: Average error distance.

detected_sources: List[str | int]

error_distances: Dict[int | str, float]

real_sources: List[str | int]

class nsdlib.common.models.SourceDetectionResult(config: nsdlib.common.models.SourceDetectionConfig, G: networkx.classes.graph.Graph, IG: networkx.classes.graph.Graph, global_scores: Dict[int | str, float], scores_in_outbreaks: List[Dict[int | str, float]], detected_sources: List[str | int])[source]

Bases: object

G: Graph

IG: Graph

config: SourceDetectionConfig

detected_sources: List[str | int]

global_scores: Dict[int | str, float]

scores_in_outbreaks: List[Dict[int | str, float]]