Skip to main content

Table 5 Comparison of various classification models for transaction categorization task

From: Deep learning enhancing banking services: a hybrid transaction classification and cash flow prediction approach

Model

Accuracy

AUC

Recall

Prec.

F1

Kappa

MCC

TT (Sec)

Extreme Gradient Boosting

0.9554

0.698

0.8012

0.9569

0.9545

0.943

0.9434

298.3

CatBoost Classifier

0.9476

0.6974

0.7398

0.9503

0.9461

0.9329

0.9335

49.64

Extra Trees Classifier

0.9457

0.6949

0.7547

0.9453

0.9447

0.9306

0.9308

15.27

Logistic Regression

0.9427

0.6969

0.7258

0.9458

0.9409

0.9267

0.9275

41.51

Gradient Boosting Classifier

0.9402

0.6944

0.7518

0.9446

0.939

0.9234

0.9243

170.4

SVM - Linear Kernel

0.9396

0

0.6948

0.9446

0.937

0.9226

0.9237

3.268

Ridge Classifier

0.937

0

0.6917

0.9437

0.9342

0.9191

0.9203

0.396

Linear Discriminant Analysis

0.934

0.6953

0.6867

0.9444

0.9333

0.9156

0.9175

8.38

K Neighbors Classifier

0.8455

0.6678

0.4984

0.8341

0.8366

0.8014

0.802

121.1

Naive Bayes

0.818

0.6401

0.7371

0.952

0.8615

0.7776

0.7866

1.359

Ada Boost Classifier

0.7951

0.5463

0.4443

0.6992

0.7353

0.7255

0.7456

3.36

Light Gradient Boosting Machine

0.6293

0.5559

0.3534

0.6996

0.6429

0.5442

0.5484

20.12

Quadratic Discriminant Analysis

0.0111

0.354

0.0496

0.2889

0.0203

0.0068

0.031

4.846

  1. Results 10-fold Cross-validation (with bold noted the proposed model)