Skip to content

API reference — Anomaly & density trees

Tree-based anomaly detection, density estimation, streaming mass scoring, collusive displacement (CoDisp), active human feedback, and contextual surprise.

GritBot

pythonclass GritBot((min_items=None, max_norm=None, sig_threshold=None, max_depth=None, check_cols=None))

GritBot.scan

pythonscan((self, /, x, col_names=None, dtypes=None))

DensityEstimationTree

pythonclass DensityEstimationTree((max_depth=8, min_samples_split=10, min_leaf_samples=5, min_density_gain=0.0, alpha=0.0))

DensityEstimationTree.fit

pythonfit((self, /, x, col_names=None, dtypes=None))

DensityEstimationTree.predict_anomaly_score

pythonpredict_anomaly_score((self, /, x))

DensityEstimationTree.predict_density

pythonpredict_density((self, /, x))

ExtendedIsolationTree

pythonclass ExtendedIsolationTree((max_depth=None, extension_level=None, subsample_size=None, seed=42))

ExtendedIsolationTree.anomaly_score

pythonanomaly_score((self, /, x))

ExtendedIsolationTree.fit

pythonfit((self, /, x, col_names=None, dtypes=None))

ExtendedIsolationTree.path_length

pythonpath_length((self, /, x))

HalfSpaceTree

pythonclass HalfSpaceTree((max_depth=6, window_size=250, initial_mass=0.0, seed=42))

HalfSpaceTree.anomaly_score

pythonanomaly_score((self, /, data))

HalfSpaceTree.fit

pythonfit((self, /, data))

HalfSpaceTree.insert

pythoninsert((self, /, point))

HalfSpaceTree.mass_score

pythonmass_score((self, /, data))

RandomCutTree

pythonclass RandomCutTree((sample_size=256, seed=42))

RandomCutTree.codisp

pythoncodisp((self, /, data))

RandomCutTree.fit

pythonfit((self, /, data))

ActiveAnomalyTree

pythonclass ActiveAnomalyTree((max_depth=8, eta=0.1, prior_weight=1.0, seed=42))

ActiveAnomalyTree.anomaly_score

pythonanomaly_score((self, /, data))

ActiveAnomalyTree.fit

pythonfit((self, /, data))

ActiveAnomalyTree.update_feedback

pythonupdate_feedback((self, /, point, is_anomaly, eta=0.5))

ContextAnomalyTree

pythonclass ContextAnomalyTree((max_depth=5, alphabet_size=256, smoothing_discount=0.5))

ContextAnomalyTree.conditional_prob

pythonconditional_prob((self, /, context, symbol))

ContextAnomalyTree.fit

pythonfit((self, /, sequences))

ContextAnomalyTree.insert

pythoninsert((self, /, context, symbol))

ContextAnomalyTree.surprise_score

pythonsurprise_score((self, /, context, symbol))