Skip to main content

Table 3 Comparison of loss, accuracy, and time of OTS and characteristic subsequences on ResNet

From: Time series classification based on statistical features

Class

Sub_acc

Ori_acc

Sub_loss

Ori_loss

Sub_time (ms)

Ori_time (ms)

MPCE

Sub

Ori

Beef

0.20000000

0.20000000

1.74496531

1.73106301

5158

3086

0.0266

0.0266

_ECG_torso

0.25000000

0.57500000

0.39711393

0.76636980

8006

12008

0.0187

0.0106

FISH

0.29142857

0.12000000

1.74508711

2.00773776

5028

6034

0.0040

0.0050

Haptics

0.21935484

0.11612903

1.59135229

1.63156084

80255

9029

0.0050

0.0057

nlineSkate

0.18000000

0.18000000

1.80085803

1.99357187

13023

13023

0.0082

0.0082

Lighting2

0.66666666

0.66666665

0.82555122

0.80486704

10167

11185

0.0055

0.0055

MALLAT

0.81818181

0.25454545

0.45194966

0.68730921

16007

22009

0.0033

0.0135

CG_Thorax1

0.09777778

0.08388889

3.28499700

3.29407895

18009

23012

0.0005

0.0005

CG_Thorax2

0.13000000

0.02277778

3.31574941

3.43187079

20010

25013

0.0005

0.0005

OliveOil

0.16666667

0.43333334

1.71029079

1.31717467

18609

20650

0.0278

0.0189

OSULeaf

0.27000000

0.22000000

1.52914039

1.45554120

21089

22093

0.0036

0.0039

ightCurves

0.96600000

0.84100000

0.11879947

0.24178773

36004

555007

0.0000

0.0001

Symbols

0.87999999

0.40000000

0.43127703

0.52733892

27027

30030

0.0048

0.0240

yoga

0.45666666

0.45666666

0.52088028

0.64692467

32011

36012

0.0018

0.0018

mean

0.399482

0.326429

1.390572

1.466943

22171.64

56299.36

0.0079

0.0089

  1. Data in italic means better experimental results