Table 7 Results of the KM-DBSCAN data reduction compared to other methods from the related work on the Adult9a dataset.

From: KM-DBSCAN: an enhanced density and centroid based border detection framework for data reduction towards green AI

Method

Accuracy

Loss/Gain

in Accuracy

Reduction

Ratio

SVs

Reduction Ratio

Speedup

Whole dataset

0.85056

–

–

–

–

FIFDR9

0.84900

− 0.00156

0.5230

0.97400

2.100

Gaffari13

0.84500

− 0.00556

0.1470

0.08000

61.890

Shell Extraction (SE)12

0.70300

− 0.14756

0.3740

0.36100

7.500

DBI14

0.84947

− 0.00109

0.9000

0.97895

1.1748

BRI14

0.84474

− 0.00582

0.4000

0.68154

3.2225

BRIX14

0.84044

− 0.01012

0.0500

0.02728

790.5368

SVO14

0.80839

− 0.04217

0.4592

0.86757

3.13133

SVOX14

0.84352

− 0.00704

0.2230

0.24568

25.380

Proposed method

0.84079

− 0.00977

0.02923

0.01629

6907.340