Evaluations

Usage

from causallearn.graph.ArrowConfusion import ArrowConfusion
from causallearn.graph.AdjacencyConfusion import AdjacencyConfusion
from causallearn.graph.SHD import SHD

# For arrows
arrow = ArrowConfusion(truth_cpdag, est)

arrowsTp = arrow.get_arrows_tp()
arrowsFp = arrow.get_arrows_fp()
arrowsFn = arrow.get_arrows_fn()
arrowsTn = arrow.get_arrows_tn()

arrowPrec = arrow.get_arrows_precision()
arrowRec = arrow.get_arrows_recall()

# For adjacency matrices
adj = AdjacencyConfusion(truth_cpdag, est)

adjTp = adj.get_adj_tp()
adjFp = adj.get_adj_fp()
adjFn = adj.get_adj_fn()
adjTn = adj.get_adj_tn()

adjPrec = adj.get_adj_precision()
adjRec = adj.get_adj_recall()

# Structural Hamming Distance
shd = SHD(truth_cpdag, est).get_shd()

Parameters

X: Data with T*D dimensions.

truth_cpdag: Graph class.

est: Graph class.

Returns

arrowsTp/Fp/Fn/Tn: True positive/false positive/false negative/true negative arrows.

arrowPrec: Precision for arrows.

arrowRec: Recall for arrows.

adjTp/Fp/Fn/Tn: True positive/false positive/false negative/true negative edges.

adjPrec: Precision for the adjacency matrix.

adjRec: Recall for the adjacency matrix.

shd: Structural Hamming Distance.