update plots

aucad · Nov 24, 2023 · 0f4a372 · 0f4a372
1 parent 2e11833
commit 0f4a372
Show file tree

Hide file tree

Showing 18 changed files with 164 additions and 165 deletions.
diff --git a/README.md b/README.md
@@ -68,23 +68,18 @@ Install dependencies
 pip install -r requirements.txt --user
 ```
 
-![duration](https://img.shields.io/badge/%F0%9F%95%92%2024%E2%80%9448%20hours-FFFF00?style=flat-square) **Run attacks.**   
-Run experiments with validation, for all combinations of data sets, classifiers, and attacks. 
-
-```
-make attacks
-```
-
-![duration](https://img.shields.io/badge/%F0%9F%95%92%2024%E2%80%9448%20hours-FFFF00?style=flat-square) **Run alternative experiments.**   
-Run experiment variants to compare with the above attacks. 
+![duration](https://img.shields.io/badge/%F0%9F%95%92%2024%E2%80%9448%20hours/each-FFFF00?style=flat-square) **Run attacks evaluations.**   
+Run experiments for combinations of data sets x classifiers x attacks (24 experiment cases). 
 
 <pre>
+make attacks   -- run all attacks, using constraint enforcement
 make original  -- run all attacks, but without validation (ignore constraints).
 make reset     -- run all attacks, but use naive reset strategy.
 </pre>
 
-![duration](https://img.shields.io/badge/%F0%9F%95%92%2030%20min%20%E2%80%94%203%20h-FFFF00?style=flat-square) **Run performance test.**   
-Uses varying number of constraints on same data set, to measure performance impact on current hardware. 
+![duration](https://img.shields.io/badge/%F0%9F%95%92%2030%20min%20%E2%80%94%203%20h-FFFF00?style=flat-square) **Run constraint performance test.**   
+Uses varying number of constraints to measure performance impact on current hardware. 
+This experiment tests PGD - C-PGD - VPGD only on same data set, UNSW-NB15.
 
 ```
 make perf

diff --git a/plot/bar.py b/plot/bar.py
@@ -40,7 +40,9 @@ def get_color_scheme(n):
 
 
 def multi_bar(ax, results, cat_names, colors):
-    llbl, rlbl = list(zip(*[(r, l) for (l, r), _ in results]))
+    llbl, rlbl = list(zip(*[
+        (r, l.replace("0", "").replace("PT-1-", "")
+         .replace("PT-2-", "")) for (l, r), _ in results]))
     uniq_rl = [x for k, v in groupby(rlbl)
                for x in [k] + [' '] * (sum(1 for __ in v) - 1)]
     labels = [i for i, _ in enumerate(results)]
@@ -114,7 +116,8 @@ def plot_acc(input_data, plot_name, data_labels,
 
     # draw sub plots
     for i, ckey in enumerate(subplots):
-        cdata = [(x[0][1:], x[1]) for x in data.values()
+        cdata = [(x[0][1:], x[1])
+                 for x in data.values()
                  if ckey == x[0][0]]
         if sort_key is not None:
             cdata.sort(key=sort_key)

diff --git a/plot/table.py b/plot/table.py
@@ -83,10 +83,11 @@ def duration_plot(self, baseline: ResultData):
 
         def fmt(r):
             c, n, a = self.cls(r), self.name(r), self.attack(r)
-            z, t, s = -1, r['experiment']['duration_sec'], '-'
+            k = r['experiment']['k_folds']
+            z, t, s = -1, r['experiment']['duration_sec']/k, '-'
             ba = a[1:] if (a[0] == 'V' or a == 'CPGD') else a
             if b := baseline.find(c, n, ba):
-                z = b['experiment']['duration_sec']
+                z = b['experiment']['duration_sec']/k
                 s = f"{t / z:.2f}" if z != 0 else s
             return [c, n, a, f"{z:.0f}", f"{t:.0f}", s]
 

diff --git a/ref_result/__bar_acc_PT-1_ref_result_perf_darwin.pdf b/ref_result/__bar_acc_PT-1_ref_result_perf_darwin.pdf
diff --git a/ref_result/__bar_acc_PT-1_ref_result_perf_debian.pdf b/ref_result/__bar_acc_PT-1_ref_result_perf_debian.pdf
diff --git a/ref_result/__bar_acc_PT-1_ref_result_perf_ubuntu.pdf b/ref_result/__bar_acc_PT-1_ref_result_perf_ubuntu.pdf
diff --git a/ref_result/__bar_acc_PT-1_ref_result_perf_windows.pdf b/ref_result/__bar_acc_PT-1_ref_result_perf_windows.pdf
diff --git a/ref_result/__bar_acc_PT-2_ref_result_perf_darwin.pdf b/ref_result/__bar_acc_PT-2_ref_result_perf_darwin.pdf
diff --git a/ref_result/__bar_acc_PT-2_ref_result_perf_debian.pdf b/ref_result/__bar_acc_PT-2_ref_result_perf_debian.pdf
diff --git a/ref_result/__bar_acc_PT-2_ref_result_perf_ubuntu.pdf b/ref_result/__bar_acc_PT-2_ref_result_perf_ubuntu.pdf
diff --git a/ref_result/__bar_acc_PT-2_ref_result_perf_windows.pdf b/ref_result/__bar_acc_PT-2_ref_result_perf_windows.pdf
diff --git a/ref_result/__time_ref_result_attacks.txt b/ref_result/__time_ref_result_attacks.txt
@@ -1,25 +1,25 @@
-  classifier    exp-name   attack   t0    t(s)   t-ratio
-Neural Network  IoT-23     CPGD      354    817     2.31
-Neural Network  IoT-23     VHSJ     3720   3492     0.94
-Neural Network  IoT-23     VPGD      354    991     2.80
-Neural Network  IoT-23     VZoo    21105  20157     0.96
-Neural Network  LCLD       CPGD      534   3157     5.92
-Neural Network  LCLD       VHSJ     7601   6998     0.92
-Neural Network  LCLD       VPGD      534   1324     2.48
-Neural Network  LCLD       VZoo    30508  28594     0.94
-Neural Network  UNSW-NB15  CPGD      315    814     2.59
-Neural Network  UNSW-NB15  VHSJ     6185   5870     0.95
-Neural Network  UNSW-NB15  VPGD      315    493     1.57
-Neural Network  UNSW-NB15  VZoo    26665  24638     0.92
-Neural Network  URL        CPGD      384   1808     4.70
-Neural Network  URL        VHSJ     3487   3232     0.93
-Neural Network  URL        VPGD      384    934     2.43
-Neural Network  URL        VZoo    17534  16557     0.94
-XGBoost         IoT-23     VHSJ      569    559     0.98
-XGBoost         IoT-23     VZoo     1110   1203     1.08
-XGBoost         LCLD       VHSJ     1355   1328     0.98
-XGBoost         LCLD       VZoo     2213   2394     1.08
-XGBoost         UNSW-NB15  VHSJ      650    648     1.00
-XGBoost         UNSW-NB15  VZoo     1216   1270     1.04
-XGBoost         URL        VHSJ      808    799     0.99
-XGBoost         URL        VZoo     1504   1625     1.08
+  classifier    exp-name   attack   t0   t(s)  t-ratio
+Neural Network  IoT-23     CPGD      71   163     2.31
+Neural Network  IoT-23     VHSJ     744   698     0.94
+Neural Network  IoT-23     VPGD      71   198     2.80
+Neural Network  IoT-23     VZoo    4221  4031     0.96
+Neural Network  LCLD       CPGD     107   631     5.92
+Neural Network  LCLD       VHSJ    1520  1400     0.92
+Neural Network  LCLD       VPGD     107   265     2.48
+Neural Network  LCLD       VZoo    6102  5719     0.94
+Neural Network  UNSW-NB15  CPGD      63   163     2.59
+Neural Network  UNSW-NB15  VHSJ    1237  1174     0.95
+Neural Network  UNSW-NB15  VPGD      63    99     1.57
+Neural Network  UNSW-NB15  VZoo    5333  4928     0.92
+Neural Network  URL        CPGD      77   362     4.70
+Neural Network  URL        VHSJ     697   646     0.93
+Neural Network  URL        VPGD      77   187     2.43
+Neural Network  URL        VZoo    3507  3311     0.94
+XGBoost         IoT-23     VHSJ     114   112     0.98
+XGBoost         IoT-23     VZoo     222   241     1.08
+XGBoost         LCLD       VHSJ     271   266     0.98
+XGBoost         LCLD       VZoo     443   479     1.08
+XGBoost         UNSW-NB15  VHSJ     130   130     1.00
+XGBoost         UNSW-NB15  VZoo     243   254     1.04
+XGBoost         URL        VHSJ     162   160     0.99
+XGBoost         URL        VZoo     301   325     1.08
diff --git a/ref_result/__time_ref_result_original.txt b/ref_result/__time_ref_result_original.txt
@@ -1,25 +1,25 @@
-  classifier    exp-name   attack   t0    t(s)   t-ratio
-Neural Network  IoT-23     CPGD      354    351     0.99
-Neural Network  IoT-23     HSJ      3720   3720     1.00
-Neural Network  IoT-23     PGD       354    354     1.00
-Neural Network  IoT-23     Zoo     21105  21105     1.00
-Neural Network  LCLD       CPGD      534    911     1.71
-Neural Network  LCLD       HSJ      7601   7601     1.00
-Neural Network  LCLD       PGD       534    534     1.00
-Neural Network  LCLD       Zoo     30508  30508     1.00
-Neural Network  UNSW-NB15  CPGD      315    499     1.59
-Neural Network  UNSW-NB15  HSJ      6185   6185     1.00
-Neural Network  UNSW-NB15  PGD       315    315     1.00
-Neural Network  UNSW-NB15  Zoo     26665  26665     1.00
-Neural Network  URL        CPGD      384    509     1.32
-Neural Network  URL        HSJ      3487   3487     1.00
-Neural Network  URL        PGD       384    384     1.00
-Neural Network  URL        Zoo     17534  17534     1.00
-XGBoost         IoT-23     HSJ       569    569     1.00
-XGBoost         IoT-23     Zoo      1110   1110     1.00
-XGBoost         LCLD       HSJ      1355   1355     1.00
-XGBoost         LCLD       Zoo      2213   2213     1.00
-XGBoost         UNSW-NB15  HSJ       650    650     1.00
-XGBoost         UNSW-NB15  Zoo      1216   1216     1.00
-XGBoost         URL        HSJ       808    808     1.00
-XGBoost         URL        Zoo      1504   1504     1.00
+  classifier    exp-name   attack   t0   t(s)  t-ratio
+Neural Network  IoT-23     CPGD      71    70     0.99
+Neural Network  IoT-23     HSJ      744   744     1.00
+Neural Network  IoT-23     PGD       71    71     1.00
+Neural Network  IoT-23     Zoo     4221  4221     1.00
+Neural Network  LCLD       CPGD     107   182     1.71
+Neural Network  LCLD       HSJ     1520  1520     1.00
+Neural Network  LCLD       PGD      107   107     1.00
+Neural Network  LCLD       Zoo     6102  6102     1.00
+Neural Network  UNSW-NB15  CPGD      63   100     1.59
+Neural Network  UNSW-NB15  HSJ     1237  1237     1.00
+Neural Network  UNSW-NB15  PGD       63    63     1.00
+Neural Network  UNSW-NB15  Zoo     5333  5333     1.00
+Neural Network  URL        CPGD      77   102     1.32
+Neural Network  URL        HSJ      697   697     1.00
+Neural Network  URL        PGD       77    77     1.00
+Neural Network  URL        Zoo     3507  3507     1.00
+XGBoost         IoT-23     HSJ      114   114     1.00
+XGBoost         IoT-23     Zoo      222   222     1.00
+XGBoost         LCLD       HSJ      271   271     1.00
+XGBoost         LCLD       Zoo      443   443     1.00
+XGBoost         UNSW-NB15  HSJ      130   130     1.00
+XGBoost         UNSW-NB15  Zoo      243   243     1.00
+XGBoost         URL        HSJ      162   162     1.00
+XGBoost         URL        Zoo      301   301     1.00
diff --git a/ref_result/__time_ref_result_perf_darwin.txt b/ref_result/__time_ref_result_perf_darwin.txt
@@ -1,19 +1,19 @@
   classifier    exp-name  attack  t0  t(s)  t-ratio
-Neural Network  PT-1-03   PGD     83    83     1.00
-Neural Network  PT-1-03   CPGD    83   142     1.70
-Neural Network  PT-1-03   VPGD    83   140     1.68
-Neural Network  PT-1-06   PGD     91    91     1.00
-Neural Network  PT-1-06   VPGD    91   165     1.82
-Neural Network  PT-1-06   CPGD    91   176     1.93
-Neural Network  PT-1-12   PGD     91    91     1.00
-Neural Network  PT-1-12   CPGD    91   245     2.70
-Neural Network  PT-1-12   VPGD    91   229     2.52
-Neural Network  PT-2-03   PGD     93    93     1.00
-Neural Network  PT-2-03   CPGD    93   152     1.64
-Neural Network  PT-2-03   VPGD    93   141     1.52
-Neural Network  PT-2-06   PGD     85    85     1.00
-Neural Network  PT-2-06   VPGD    85   171     2.01
-Neural Network  PT-2-06   CPGD    85   181     2.12
-Neural Network  PT-2-12   PGD     87    87     1.00
-Neural Network  PT-2-12   CPGD    87   238     2.73
-Neural Network  PT-2-12   VPGD    87   232     2.66
+Neural Network  PT-1-03   PGD     17    17     1.00
+Neural Network  PT-1-03   CPGD    17    28     1.70
+Neural Network  PT-1-03   VPGD    17    28     1.68
+Neural Network  PT-1-06   PGD     18    18     1.00
+Neural Network  PT-1-06   VPGD    18    33     1.82
+Neural Network  PT-1-06   CPGD    18    35     1.93
+Neural Network  PT-1-12   PGD     18    18     1.00
+Neural Network  PT-1-12   CPGD    18    49     2.70
+Neural Network  PT-1-12   VPGD    18    46     2.52
+Neural Network  PT-2-03   PGD     19    19     1.00
+Neural Network  PT-2-03   CPGD    19    30     1.64
+Neural Network  PT-2-03   VPGD    19    28     1.52
+Neural Network  PT-2-06   PGD     17    17     1.00
+Neural Network  PT-2-06   VPGD    17    34     2.01
+Neural Network  PT-2-06   CPGD    17    36     2.12
+Neural Network  PT-2-12   PGD     17    17     1.00
+Neural Network  PT-2-12   CPGD    17    48     2.73
+Neural Network  PT-2-12   VPGD    17    46     2.66
diff --git a/ref_result/__time_ref_result_perf_debian.txt b/ref_result/__time_ref_result_perf_debian.txt
@@ -1,19 +1,19 @@
-  classifier    exp-name  attack  t0   t(s)  t-ratio
-Neural Network  PT-1-03   PGD     328   328     1.00
-Neural Network  PT-1-03   VPGD    328   474     1.44
-Neural Network  PT-1-03   CPGD    328   659     2.01
-Neural Network  PT-1-06   PGD     328   328     1.00
-Neural Network  PT-1-06   VPGD    328   560     1.71
-Neural Network  PT-1-06   CPGD    328   812     2.48
-Neural Network  PT-1-12   PGD     372   372     1.00
-Neural Network  PT-1-12   CPGD    372  1072     2.89
-Neural Network  PT-1-12   VPGD    372   731     1.97
-Neural Network  PT-2-03   PGD     344   344     1.00
-Neural Network  PT-2-03   VPGD    344   466     1.35
-Neural Network  PT-2-03   CPGD    344   690     2.01
-Neural Network  PT-2-06   PGD     323   323     1.00
-Neural Network  PT-2-06   VPGD    323   549     1.70
-Neural Network  PT-2-06   CPGD    323   780     2.41
-Neural Network  PT-2-12   PGD     323   323     1.00
-Neural Network  PT-2-12   VPGD    323   721     2.23
-Neural Network  PT-2-12   CPGD    323  1056     3.27
+  classifier    exp-name  attack  t0  t(s)  t-ratio
+Neural Network  PT-1-03   PGD     66    66     1.00
+Neural Network  PT-1-03   VPGD    66    95     1.44
+Neural Network  PT-1-03   CPGD    66   132     2.01
+Neural Network  PT-1-06   PGD     66    66     1.00
+Neural Network  PT-1-06   VPGD    66   112     1.71
+Neural Network  PT-1-06   CPGD    66   162     2.48
+Neural Network  PT-1-12   PGD     74    74     1.00
+Neural Network  PT-1-12   CPGD    74   214     2.89
+Neural Network  PT-1-12   VPGD    74   146     1.97
+Neural Network  PT-2-03   PGD     69    69     1.00
+Neural Network  PT-2-03   VPGD    69    93     1.35
+Neural Network  PT-2-03   CPGD    69   138     2.01
+Neural Network  PT-2-06   PGD     65    65     1.00
+Neural Network  PT-2-06   VPGD    65   110     1.70
+Neural Network  PT-2-06   CPGD    65   156     2.41
+Neural Network  PT-2-12   PGD     65    65     1.00
+Neural Network  PT-2-12   VPGD    65   144     2.23
+Neural Network  PT-2-12   CPGD    65   211     3.27
diff --git a/ref_result/__time_ref_result_perf_ubuntu.txt b/ref_result/__time_ref_result_perf_ubuntu.txt
@@ -1,19 +1,19 @@
-  classifier    exp-name  attack  t0   t(s)  t-ratio
-Neural Network  PT-1-03   PGD     284   284     1.00
-Neural Network  PT-1-03   VPGD    284   428     1.51
-Neural Network  PT-1-03   CPGD    284   566     1.99
-Neural Network  PT-1-06   PGD     276   276     1.00
-Neural Network  PT-1-06   VPGD    276   479     1.73
-Neural Network  PT-1-06   CPGD    276   766     2.77
-Neural Network  PT-1-12   PGD     243   243     1.00
-Neural Network  PT-1-12   CPGD    243   978     4.02
-Neural Network  PT-1-12   VPGD    243   648     2.67
-Neural Network  PT-2-03   PGD     259   259     1.00
-Neural Network  PT-2-03   CPGD    259   624     2.40
-Neural Network  PT-2-03   VPGD    259   408     1.57
-Neural Network  PT-2-06   PGD     287   287     1.00
-Neural Network  PT-2-06   CPGD    287   724     2.52
-Neural Network  PT-2-06   VPGD    287   492     1.71
-Neural Network  PT-2-12   PGD     267   267     1.00
-Neural Network  PT-2-12   VPGD    267   657     2.46
-Neural Network  PT-2-12   CPGD    267   948     3.56
+  classifier    exp-name  attack  t0  t(s)  t-ratio
+Neural Network  PT-1-03   PGD     57    57     1.00
+Neural Network  PT-1-03   VPGD    57    86     1.51
+Neural Network  PT-1-03   CPGD    57   113     1.99
+Neural Network  PT-1-06   PGD     55    55     1.00
+Neural Network  PT-1-06   VPGD    55    96     1.73
+Neural Network  PT-1-06   CPGD    55   153     2.77
+Neural Network  PT-1-12   PGD     49    49     1.00
+Neural Network  PT-1-12   CPGD    49   196     4.02
+Neural Network  PT-1-12   VPGD    49   130     2.67
+Neural Network  PT-2-03   PGD     52    52     1.00
+Neural Network  PT-2-03   CPGD    52   125     2.40
+Neural Network  PT-2-03   VPGD    52    82     1.57
+Neural Network  PT-2-06   PGD     57    57     1.00
+Neural Network  PT-2-06   CPGD    57   145     2.52
+Neural Network  PT-2-06   VPGD    57    98     1.71
+Neural Network  PT-2-12   PGD     53    53     1.00
+Neural Network  PT-2-12   VPGD    53   131     2.46
+Neural Network  PT-2-12   CPGD    53   190     3.56
diff --git a/ref_result/__time_ref_result_perf_windows.txt b/ref_result/__time_ref_result_perf_windows.txt
@@ -1,19 +1,19 @@
-  classifier    exp-name  attack  t0   t(s)  t-ratio
-Neural Network  PT-1-03   PGD      99    99     1.00
-Neural Network  PT-1-03   VPGD     99   175     1.76
-Neural Network  PT-1-03   CPGD     99   210     2.11
-Neural Network  PT-1-06   PGD     104   104     1.00
-Neural Network  PT-1-06   CPGD    104   271     2.61
-Neural Network  PT-1-06   VPGD    104   227     2.18
-Neural Network  PT-1-12   PGD      95    95     1.00
-Neural Network  PT-1-12   CPGD     95   365     3.83
-Neural Network  PT-1-12   VPGD     95   304     3.19
-Neural Network  PT-2-03   PGD     101   101     1.00
-Neural Network  PT-2-03   CPGD    101   224     2.22
-Neural Network  PT-2-03   VPGD    101   190     1.88
-Neural Network  PT-2-06   PGD     104   104     1.00
-Neural Network  PT-2-06   CPGD    104   267     2.58
-Neural Network  PT-2-06   VPGD    104   232     2.24
-Neural Network  PT-2-12   PGD     104   104     1.00
-Neural Network  PT-2-12   VPGD    104   307     2.95
-Neural Network  PT-2-12   CPGD    104   358     3.45
+  classifier    exp-name  attack  t0  t(s)  t-ratio
+Neural Network  PT-1-03   PGD     20    20     1.00
+Neural Network  PT-1-03   VPGD    20    35     1.76
+Neural Network  PT-1-03   CPGD    20    42     2.11
+Neural Network  PT-1-06   PGD     21    21     1.00
+Neural Network  PT-1-06   CPGD    21    54     2.61
+Neural Network  PT-1-06   VPGD    21    45     2.18
+Neural Network  PT-1-12   PGD     19    19     1.00
+Neural Network  PT-1-12   CPGD    19    73     3.83
+Neural Network  PT-1-12   VPGD    19    61     3.19
+Neural Network  PT-2-03   PGD     20    20     1.00
+Neural Network  PT-2-03   CPGD    20    45     2.22
+Neural Network  PT-2-03   VPGD    20    38     1.88
+Neural Network  PT-2-06   PGD     21    21     1.00
+Neural Network  PT-2-06   CPGD    21    53     2.58
+Neural Network  PT-2-06   VPGD    21    46     2.24
+Neural Network  PT-2-12   PGD     21    21     1.00
+Neural Network  PT-2-12   VPGD    21    61     2.95
+Neural Network  PT-2-12   CPGD    21    72     3.45
diff --git a/ref_result/__time_ref_result_reset.txt b/ref_result/__time_ref_result_reset.txt
@@ -1,25 +1,25 @@
-  classifier    exp-name   attack   t0    t(s)   t-ratio
-Neural Network  IoT-23     CPGD      354    750     2.11
-Neural Network  IoT-23     VHSJ     3720   3583     0.96
-Neural Network  IoT-23     VPGD      354   1010     2.85
-Neural Network  IoT-23     VZoo    21105  20220     0.96
-Neural Network  LCLD       CPGD      534   3268     6.12
-Neural Network  LCLD       VHSJ     7601   7174     0.94
-Neural Network  LCLD       VPGD      534   1349     2.53
-Neural Network  LCLD       VZoo    30508  29336     0.96
-Neural Network  UNSW-NB15  CPGD      315    824     2.62
-Neural Network  UNSW-NB15  VHSJ     6185   5907     0.96
-Neural Network  UNSW-NB15  VPGD      315    524     1.66
-Neural Network  UNSW-NB15  VZoo    26665  25470     0.96
-Neural Network  URL        CPGD      384   1871     4.87
-Neural Network  URL        VHSJ     3487   3337     0.96
-Neural Network  URL        VPGD      384    930     2.42
-Neural Network  URL        VZoo    17534  17146     0.98
-XGBoost         IoT-23     VHSJ      569    564     0.99
-XGBoost         IoT-23     VZoo     1110   1192     1.07
-XGBoost         LCLD       VHSJ     1355   1340     0.99
-XGBoost         LCLD       VZoo     2213   2432     1.10
-XGBoost         UNSW-NB15  VHSJ      650    665     1.02
-XGBoost         UNSW-NB15  VZoo     1216   1312     1.08
-XGBoost         URL        VHSJ      808    827     1.02
-XGBoost         URL        VZoo     1504   1694     1.13
+  classifier    exp-name   attack   t0   t(s)  t-ratio
+Neural Network  IoT-23     CPGD      71   150     2.11
+Neural Network  IoT-23     VHSJ     744   717     0.96
+Neural Network  IoT-23     VPGD      71   202     2.85
+Neural Network  IoT-23     VZoo    4221  4044     0.96
+Neural Network  LCLD       CPGD     107   654     6.12
+Neural Network  LCLD       VHSJ    1520  1435     0.94
+Neural Network  LCLD       VPGD     107   270     2.53
+Neural Network  LCLD       VZoo    6102  5867     0.96
+Neural Network  UNSW-NB15  CPGD      63   165     2.62
+Neural Network  UNSW-NB15  VHSJ    1237  1181     0.96
+Neural Network  UNSW-NB15  VPGD      63   105     1.66
+Neural Network  UNSW-NB15  VZoo    5333  5094     0.96
+Neural Network  URL        CPGD      77   374     4.87
+Neural Network  URL        VHSJ     697   667     0.96
+Neural Network  URL        VPGD      77   186     2.42
+Neural Network  URL        VZoo    3507  3429     0.98
+XGBoost         IoT-23     VHSJ     114   113     0.99
+XGBoost         IoT-23     VZoo     222   238     1.07
+XGBoost         LCLD       VHSJ     271   268     0.99
+XGBoost         LCLD       VZoo     443   486     1.10
+XGBoost         UNSW-NB15  VHSJ     130   133     1.02
+XGBoost         UNSW-NB15  VZoo     243   262     1.08
+XGBoost         URL        VHSJ     162   165     1.02
+XGBoost         URL        VZoo     301   339     1.13