Detecting Illegal Online Gambling (IOG) Services in the Mobile Environment
Table 9
Results for comprehensive experiments.
Algorithm
KoNLPy.Option
Accuracy
Precision
Recall
F1-score
linearSVM
Okt.morphs
0.9788
0.9781
0.9791
0.9786
Okt.nouns
0.9759
0.9753
0.9760
0.9757
Okt.phrases
0.9784
0.9776
0.9788
0.9782
Mecab.morphs
0.9795
0.9791
0.9794
0.9792
Mecab.nouns
0.9757
0.9750
0.9759
0.9754
LR
Okt.morphs
0.9766
0.9756
0.9772
0.9763
Okt.nouns
0.9742
0.9736
0.9743
0.9739
Okt.phrases
0.9776
0.9768
0.9780
0.9774
Mecab.morphs
0.9780
0.9777
0.9778
0.9778
Mecab.nouns
0.9761
0.9754
0.9762
0.9758
NB
Okt.morphs
0.9750
0.9747
0.9747
0.9747
Okt.nouns
0.9699
0.9695
0.9695
0.9695
Okt.phrases
0.9712
0.9714
0.9703
0.9708
Mecab.morphs
0.9745
0.9746
0.9738
0.9742
Mecab.nouns
0.9725
0.9720
0.9724
0.9722
RF
Okt.morphs
0.9376
0.9361
0.9407
0.9373
Okt.nouns
0.9551
0.9535
0.9568
0.9548
Okt.phrases
0.9278
0.9262
0.9305
0.9274
Mecab.morphs
0.9520
0.9503
0.9544
0.9517
Mecab.nouns
0.9617
0.9602
0.9628
0.9613
SVM
Okt.morphs
0.9787
0.9781
0.9788
0.9784
Okt.nouns
0.9779
0.9773
0.9780
0.9776
Okt.phrases
0.9775
0.9771
0.9774
0.9772
Mecab.morphs
0.9785
0.9785
0.9781
0.9783
Mecab.nouns
0.9767
0.9763
0.9765
0.9764
MLP
Okt.morphs
0.9753
0.9749
0.9751
0.9750
Okt.nouns
0.9744
0.9737
0.9745
0.9741
Okt.phrases
0.9738
0.9727
0.9746
0.9736
Mecab.morphs
0.9753
0.9752
0.9747
0.9750
Mecab.nouns
0.9720
0.9713
0.9721
0.9717
Boosting (AdaBoost)
Okt.morphs
0.9640
0.9637
0.9635
0.9636
Okt.nouns
0.9729
0.9723
0.9730
0.9726
Okt.phrases
0.9687
0.9681
0.9686
0.9684
Mecab.morphs
0.9686
0.9684
0.9680
0.9682
Mecab.nouns
0.9716
0.9711
0.9715
0.9713
The F1-score metric was determined by calculating recall and precision as well as accuracy. This is the result of using the metrics.classification_report function of sklearn.
