Unsupervised Offline Changepoint Detection Ensembles

Benchmark Model Rank Results
change-point-detection-on-skabBinSeg CPD algorithm (Mahalanobis metric)#1NAB (standard): 24.1NAB (lowFP): 21.69NAB (LowFN): 25.04
change-point-detection-on-skabOptEnsemble CPDE algorithm (WeightedSum+Rank)#2NAB (standard): 23.07NAB (lowFP): 20.52NAB (LowFN): 24.35
change-point-detection-on-skabOpt CPD algorithm (Mahalanobis metric)#3NAB (standard): 22.37NAB (lowFP): 19.9NAB (LowFN): 23.37
change-point-detection-on-skabWinEnsemble CPDE algorithm (Sum+MinAbs)#4NAB (standard): 19.38NAB (lowFP): 17.03NAB (LowFN): 20.35
change-point-detection-on-skabWin CPD algorithm (l1 metric)#5NAB (standard): 18.4NAB (lowFP): 16.22NAB (LowFN): 19.19
change-point-detection-on-skabBinSegEnsemble CPDE algorithm (WeightedSum+Rank)#6NAB (standard): 18.1NAB (lowFP): 15.36NAB (LowFN): 19.51
change-point-detection-on-tepOptEnsemble CPDE algorithm (Min+MinMax/Rank)#1NAB (standard): 41.81NAB (lowFP): 41NAB (LowFN): 42.16
change-point-detection-on-tepBinSegEnsemble CPDE algorithm (Min+MinMax/Rank)#2NAB (standard): 41.81NAB (lowFP): 41NAB (LowFN): 42.16
change-point-detection-on-tepBinSeg CPD algorithm (Mahalanobis metric)#3NAB (standard): 36.88NAB (lowFP): 35.82NAB (LowFN): 37.29
change-point-detection-on-tepOpt CPD algorithm (Mahalanobis metric)#4NAB (standard): 36.88NAB (lowFP): 35.82NAB (LowFN): 37.29
change-point-detection-on-tepWin CPD algorithm (Mahalanobis metric)#5NAB (standard): 27.79NAB (lowFP): 27NAB (LowFN): 28.05
change-point-detection-on-tepWinEnsemble CPDE algorithm (WeightedSum+MinAbs)#6NAB (standard): 25.14NAB (lowFP): 24.33NAB (LowFN): 26.29