Skip to main content

Table 2 AUC values achieved by each algorithm on the Purdue and AML data sets

From: A non-parametric Bayesian model for joint cell clustering and cluster matching: identification of anomalous sample phenotypes with random effects

 

Purdue

AML tubes

  

2

3

4

5

6

7

All

ASPIRE

1.000

0.940

0.974

0.991

0.999

0.992

0.971

0.997

 

(0.000)

(0.025)

(0.003)

(0.003)

(0.001)

(0.010)

(0.005)

(0.002)

DPGMM

0.995

0.782

0.612

0.933

0.935

0.954

0.514

0.773

 

(0.010)

(0.068)

(0.070)

(0.027)

(0.019)

(0.015)

(0.120)

(0.088)

FLAME

0.930

-

-

-

-

-

-

-

 

(0.000)

       

flowPeaks

0.944

0.369

0.430

0.982

0.806

0.906

0.670

0.857

 

(0.000)

(0.003)

(0.001)

(0.001)

(0.004)

(0.002)

(0.015)

(0.038)

HDPGMM

0.576

0.452

0.493

0.530

0.600

0.571

0.509

0.532

 

(0.005)

(0.011)

(0.015)

(0.011)

(0.015)

(0.027)

(0.011)

(0.009)

  1. Numbers in parentheses are standard deviations across ten repetitions of the corresponding one-class classifiers. FLAME results for the AML data set are not included as FLAME produced errors on many of the samples in this data set.