TableĀ 1 Information about the experimental datasets before and after the preprocessing steps.

From: Attention-enhanced and integrated deep learning approach for fishing vessel classification based on multiple features

Types

Gillnetter

Hook and liner

Trawler

Fish carrier

Stow net vessel

Total

Training dataset

Quantity of trajectory before processing

2,200,392

461,421

2,263,518

783,564

977,451

6,686,346

Quantity of trajectory after processing

1,919,832

430,860

2,261,013

682,863

960,417

6,254,985

Quantity of fishing vessels after processing

3,832

860

4,513

1,363

1,917

12,485

Independent testing dataset

Quantity of trajectory before processing

110,661

27,199

113,532

45,284

48,616

306,939

Quantity of trajectory after processing

94,637

26,462

99,283

41,208

45,349

306,939

Quantity of fishing vessels after processing

200

200

200

200

200

1,000