Skip to main content

Table 8 The run-time (in HH:MM:SS format) of the AMLBID, Autosklearn and TPOT tools on the benchmark datasets

From: Using meta-learning for automated algorithms selection and configuration: an experimental framework for industrial big data

Dataset

Dataset size

AMLBID

Autosklearn

TPOT

[44]

959

00:00:05

01:23:47

00:08:14

[45]

2000

00:00:12

01:49:21

00:13:57

[46]

61000

00:05:29

04:19:05

03:42:09

[47]

274627

00:11:43

08:19:37

06:09:51

[48]

5000

00:01:27

02:31:07

01:38:36

[49]

1567

00:00:53

01:33:45

00:19:47

[50]

5388

00:00:57

01:56:50

00:55:51

[51]

1567

00:00:33

00:58:50

00:21:12

Wafer-ds

7306

00:02:17

03:44:26

01:42:21

HTRU

54641

00:06:59

03:42:09

02:57:11

vehicle

8463

00:02:28

02:12:40

01:45:40

Cnae-9

63260

00:05:47

04:07:39

03:24:52

Gas_Sens

4188

00:01:14

02:47:20

00:42:36

Covertype

25524

00:03:04

01:28:31

01:36:14

Kc1

2108

00:00:38

04:19:26

04:51:02

jannis

8641

00:01:41

02:31:07

01:41:51

MiniBooNE

52147

00:04:23

03:59:56

02:11:01

KDDCup

49402

00:05:06

03:47:20

02:37:38

segment

2310

00:00:25

01:15:45

00:33:02

Higgs

110000

00:06:16

07:37:55

05:43:24

Credi-g

30000

00:04:39

02:03:34

05:33:03

shuttle

57999

00:05:48

05:15:45

04:26:03

APS Failure

60000

00:05:39

03:58:39

05:23:35

nomao

31772

00:04:08

03:01:15

02:49:36

CustSat

76020

00:06:06

05:07:03

04:09:36

kr-vs-kp

3196

00:00:54

01:17:19

00:22:44

car

1728

00:00:38

01:38:30

00:40:07

albert

43824

00:06:27

04:09:17

03:01:03

airlines

5473

00:01:40

02:18:27

00:57:52

Numerai28.6

6574

00:03:22

02:07:39

01:16:17

  1. The best performances among all AutoML frameworks are highlighted in bold