【深度学习】CICIDS 2019，入侵检测，SVM支持向量机，随机丛林，DNN训练， ...

温锦文欧普厨电及净水器总代理  论坛元老 | 2024-6-13 20:00:35 | 显示全部楼层 | 阅读模式
打印 上一主题 下一主题
楼主

主题 1761|帖子 1761|积分 5283

数据集先容

下载：https://www.kaggle.com/datasets/tarundhamor/cicids-2019-dataset
数据个数：
  1.     # 删除label中是WebDDoS的数据
  2.     df = df[df['Label'] != 'WebDDoS']
复制代码

Python情况

  1. pandas
  2. scipy
  3. scikit-learn==1.3
  4. torch
  5. matplotlib
  6. seaborn
  7. numpy
复制代码
随机丛林训练结果

  1. 特征重要程度:
  2. Max Packet Length: 0.0684
  3. Packet Length Mean: 0.0637
  4. Avg Fwd Segment Size: 0.0579
  5. Fwd Packet Length Max: 0.0569
  6. Average Packet Size: 0.0569
  7. Subflow Fwd Bytes: 0.0522
  8. Fwd Packet Length Min: 0.0505
  9. Fwd Packet Length Mean: 0.0501
  10. Min Packet Length: 0.0474
  11. Total Length of Fwd Packets: 0.0401
  12. act_data_pkt_fwd: 0.0301
  13. ACK Flag Count: 0.0300
  14. Flow Bytes/s: 0.0263
  15. Init_Win_bytes_forward: 0.0232
  16. Inbound: 0.0224
  17. Flow IAT Max: 0.0210
  18. Flow IAT Std: 0.0193
  19. Flow IAT Min: 0.0182
  20. Fwd IAT Max: 0.0169
  21. Flow IAT Mean: 0.0162
  22. Total Fwd Packets: 0.0162
  23. Protocol: 0.0161
  24. Flow Duration: 0.0156
  25. Fwd IAT Min: 0.0155
  26. Fwd IAT Total: 0.0146
  27. Fwd IAT Std: 0.0144
  28. Subflow Fwd Packets: 0.0125
  29. Fwd IAT Mean: 0.0121
  30. Fwd Packets/s: 0.0120
  31. Flow Packets/s: 0.0115
  32. URG Flag Count: 0.0086
  33. Bwd Packets/s: 0.0083
  34. Packet Length Variance: 0.0076
  35. min_seg_size_forward: 0.0067
  36. Init_Win_bytes_backward: 0.0058
  37. Bwd Header Length: 0.0052
  38. Total Backward Packets: 0.0044
  39. Packet Length Std: 0.0044
  40. Bwd IAT Total: 0.0036
  41. Subflow Bwd Bytes: 0.0034
  42. Down/Up Ratio: 0.0028
  43. Bwd IAT Max: 0.0026
  44. Avg Bwd Segment Size: 0.0023
  45. Bwd IAT Mean: 0.0021
  46. Bwd Packet Length Max: 0.0021
  47. Fwd Packet Length Std: 0.0019
  48. Total Length of Bwd Packets: 0.0019
  49. Bwd Packet Length Mean: 0.0018
  50. Fwd Header Length.1: 0.0017
  51. CWE Flag Count: 0.0017
  52. Fwd Header Length: 0.0017
  53. Subflow Bwd Packets: 0.0016
  54. Bwd IAT Min: 0.0015
  55. Idle Std: 0.0012
  56. Bwd Packet Length Min: 0.0011
  57. Idle Max: 0.0011
  58. Active Min: 0.0008
  59. Bwd IAT Std: 0.0008
  60. Active Mean: 0.0006
  61. Idle Mean: 0.0006
  62. Idle Min: 0.0004
  63. Bwd Packet Length Std: 0.0004
  64. Fwd PSH Flags: 0.0003
  65. RST Flag Count: 0.0003
  66. Active Max: 0.0001
  67. Active Std: 0.0001
  68. SYN Flag Count: 0.0001
  69. Bwd PSH Flags: 0.0000
  70. Fwd URG Flags: 0.0000
  71. Bwd URG Flags: 0.0000
  72. FIN Flag Count: 0.0000
  73. PSH Flag Count: 0.0000
  74. ECE Flag Count: 0.0000
  75. Fwd Avg Bytes/Bulk: 0.0000
  76. Fwd Avg Packets/Bulk: 0.0000
  77. Fwd Avg Bulk Rate: 0.0000
  78. Bwd Avg Bytes/Bulk: 0.0000
  79. Bwd Avg Packets/Bulk: 0.0000
  80. Bwd Avg Bulk Rate: 0.0000
  81. SimillarHTTP: 0.0000
  82. {'BENIGN': {'precision': 0.9965305156915313, 'recall': 0.9992093611638204, 'f1-score': 0.9978681405448085, 'support': 6324.0}, 'DrDoS_DNS': {'precision': 0.8426485397784491, 'recall': 0.52296875, 'f1-score': 0.6453914384882375, 'support': 6400.0}, 'DrDoS_LDAP': {'precision': 0.4552080170057698, 'recall': 0.70265625, 'f1-score': 0.5524909392468825, 'support': 6400.0}, 'DrDoS_MSSQL': {'precision': 0.43775933609958506, 'recall': 0.0659375, 'f1-score': 0.11461162411732753, 'support': 6400.0}, 'DrDoS_NTP': {'precision': 0.9717636167258139, 'recall': 0.9840625, 'f1-score': 0.9778743886344227, 'support': 6400.0}, 'DrDoS_NetBIOS': {'precision': 0.8055891441623001, 'recall': 0.936875, 'f1-score': 0.8662862096366395, 'support': 6400.0}, 'DrDoS_SNMP': {'precision': 0.9993749023284888, 'recall': 0.99921875, 'f1-score': 0.9992968200640675, 'support': 6400.0}, 'DrDoS_SSDP': {'precision': 0.6196612283071807, 'recall': 0.93171875, 'f1-score': 0.744305061474131, 'support': 6400.0}, 'Syn': {'precision': 0.9995313964386129, 'recall': 0.99984375, 'f1-score': 0.9996875488204968, 'support': 6400.0}, 'UDP-lag': {'precision': 0.6013483146067415, 'recall': 0.418125, 'f1-score': 0.4932718894009216, 'support': 6400.0},
  83. 'lDos': {'precision': 0.5524492234169653, 'recall': 0.7225, 'f1-score': 0.6261340555179418, 'support': 6400.0}, 'accuracy': 0.7527444400204767, 'macro avg': {'precision': 0.7528967485964944, 'recall': 0.7530105101058019, 'f1-score': 0.7288380105405342, 'support': 70324.0}, 'weighted avg': {'precision': 0.7526334506285287, 'recall': 0.7527444400204767, 'f1-score': 0.7285472664150532, 'support': 70324.0}}
  84. 整体准确率: 0.75
  85. 平均准确率: 0.75, 平均召回率: 0.75
  87. Process finished with exit code 0
复制代码
SVM支持向量机训练结果

  2. {'BENIGN': {'precision': 0.9767441860465116, 'recall': 0.6402439024390244, 'f1-score': 0.7734806629834254, 'support': 656.0}, 'DrDoS_DNS': {'precision': 0.7, 'recall': 0.010869565217391304, 'f1-score': 0.021406727828746176, 'support': 644.0}, 'DrDoS_LDAP': {'precision': 0.31919406150583246, 'recall': 0.9525316455696202, 'f1-score': 0.4781572676727562, 'support': 632.0}, 'DrDoS_MSSQL': {'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 652.0}, 'DrDoS_NTP': {'precision': 0.688212927756654, 'recall': 0.8418604651162791, 'f1-score': 0.7573221757322176, 'support': 645.0}, 'DrDoS_NetBIOS': {'precision': 0.7403100775193798, 'recall': 0.8995290423861853, 'f1-score': 0.81218993621545, 'support': 637.0}, 'DrDoS_SNMP': {'precision': 0.9861111111111112, 'recall': 1.0, 'f1-score': 0.993006993006993, 'support': 639.0}, 'DrDoS_SSDP': {'precision': 0.6170212765957447, 'recall': 0.23311897106109325, 'f1-score': 0.338389731621937, 'support': 622.0}, 'Syn': {'precision': 0.6777301927194861, 'recall': 1.0, 'f1-score': 0.8079132099553287, 'support': 633.0}, 'UDP-lag': {'precision': 0.4148148148148148, 'recall': 0.17582417582417584, 'f1-score': 0.24696802646086002, 'support': 637.0},
  3. 'lDos': {'precision': 0.4900662251655629, 'recall': 0.8144654088050315, 'f1-score': 0.61193148257531, 'support': 636.0}, 'accuracy': 0.5960472060287217, 'macro avg': {'precision': 0.6009277157486452, 'recall': 0.5971311978562547, 'f1-score': 0.5309787467320931, 'support': 7033.0}, 'weighted avg': {'precision': 0.6011068239694879, 'recall': 0.5960472060287217, 'f1-score': 0.5306623678089073, 'support': 7033.0}}
  4. 整体准确率: 0.60
  5. 平均准确率: 0.60, 平均召回率: 0.60
  6. ==============================
  7. svm_classification
  9. Macro-average Precision: 0.6014285714285714
  10. Macro-average Recall: 0.5971428571428571
  12. Process finished with exit code 0
复制代码
DNN训练结果

  1. C:\ProgramData\miniconda3\envs\dlib_align\python.exe C:\Users\Administrator\PycharmProjects\pythonProject3\cicids2019\x07_0DNN训练.py
  2. 训练数据的总数据形状 torch.Size([281292, 56])
  3. 训练数据的标签形状 torch.Size([281292, 11])
  4. 测试数据的总数据形状 torch.Size([70324, 56])
  5. 测试数据的标签形状 torch.Size([70324, 11])
  6. Training Progress:   3%|▎         | 1/30 [00:23<11:09, 23.10s/it]Epoch 1/30, Loss: 0.7203, Accuracy: 0.70
  7. Training Progress:   7%|▋         | 2/30 [00:45<10:29, 22.48s/it]Epoch 2/30, Loss: 0.6408, Accuracy: 0.70
  8. Epoch 3/30, Loss: 0.6217, Accuracy: 0.71
  9. Training Progress:  13%|█▎        | 4/30 [01:29<09:36, 22.18s/it]Epoch 4/30, Loss: 0.6075, Accuracy: 0.72
  10. Epoch 5/30, Loss: 0.5966, Accuracy: 0.72
  11. Training Progress:  20%|██        | 6/30 [02:13<08:49, 22.05s/it]Epoch 6/30, Loss: 0.5975, Accuracy: 0.72
  12. Training Progress:  23%|██▎       | 7/30 [02:34<08:25, 22.00s/it]Epoch 7/30, Loss: 0.5947, Accuracy: 0.72
  13. Epoch 8/30, Loss: 0.6020, Accuracy: 0.72
  14. Training Progress:  30%|███       | 9/30 [03:18<07:41, 21.99s/it]Epoch 9/30, Loss: 0.6100, Accuracy: 0.72
  15. Training Progress:  33%|███▎      | 10/30 [03:40<07:18, 21.90s/it]Epoch 10/30, Loss: 0.6152, Accuracy: 0.72
  16. Training Progress:  37%|███▋      | 11/30 [04:03<06:59, 22.10s/it]Epoch 11/30, Loss: 0.6161, Accuracy: 0.71
  17. Epoch 12/30, Loss: 0.6089, Accuracy: 0.72
  18. Training Progress:  43%|████▎     | 13/30 [04:45<06:06, 21.54s/it]Epoch 13/30, Loss: 0.6007, Accuracy: 0.71
  19. Epoch 14/30, Loss: 0.5893, Accuracy: 0.72
  20. Training Progress:  50%|█████     | 15/30 [05:27<05:20, 21.34s/it]Epoch 15/30, Loss: 0.5800, Accuracy: 0.73
  21. Training Progress:  53%|█████▎    | 16/30 [05:48<04:59, 21.38s/it]Epoch 16/30, Loss: 0.5779, Accuracy: 0.73
  22. Training Progress:  57%|█████▋    | 17/30 [06:09<04:34, 21.15s/it]Epoch 17/30, Loss: 0.5786, Accuracy: 0.72
  23. Training Progress:  60%|██████    | 18/30 [06:30<04:12, 21.02s/it]Epoch 18/30, Loss: 0.5861, Accuracy: 0.72
  24. Training Progress:  63%|██████▎   | 19/30 [06:50<03:50, 20.92s/it]Epoch 19/30, Loss: 0.5963, Accuracy: 0.72
  25. Epoch 20/30, Loss: 0.6042, Accuracy: 0.71
  26. Training Progress:  70%|███████   | 21/30 [07:30<03:03, 20.43s/it]Epoch 21/30, Loss: 0.6054, Accuracy: 0.71
  27. Epoch 22/30, Loss: 0.6027, Accuracy: 0.72
  28. Training Progress:  77%|███████▋  | 23/30 [08:11<02:23, 20.44s/it]Epoch 23/30, Loss: 0.5938, Accuracy: 0.72
  29. Training Progress:  80%|████████  | 24/30 [08:32<02:02, 20.42s/it]Epoch 24/30, Loss: 0.5836, Accuracy: 0.72
  30. Epoch 25/30, Loss: 0.5760, Accuracy: 0.73
  31. Training Progress:  87%|████████▋ | 26/30 [09:12<01:21, 20.39s/it]Epoch 26/30, Loss: 0.5753, Accuracy: 0.73
  32. Epoch 27/30, Loss: 0.5745, Accuracy: 0.73
  33. Training Progress:  93%|█████████▎| 28/30 [09:53<00:40, 20.43s/it]Epoch 28/30, Loss: 0.5813, Accuracy: 0.73
  34. Epoch 29/30, Loss: 0.5922, Accuracy: 0.70
  35. Training Progress: 100%|██████████| 30/30 [10:34<00:00, 21.15s/it]
  36. Epoch 30/30, Loss: 0.5993, Accuracy: 0.71
  37. {'0': {'precision': 0.988443881589362, 'recall': 0.9873497786211258, 'f1-score': 0.987896527173483, 'support': 6324.0}, '1': {'precision': 0.6740442655935613, 'recall': 0.41875, 'f1-score': 0.5165767154973014, 'support': 6400.0}, '2': {'precision': 0.42294757665677546, 'recall': 0.668125, 'f1-score': 0.5179890975166566, 'support': 6400.0}, '3': {'precision': 0.4377224199288256, 'recall': 0.0384375, 'f1-score': 0.07066934788853778, 'support': 6400.0}, '4': {'precision': 0.9512785072563925, 'recall': 0.8603125, 'f1-score': 0.9035116508040696, 'support': 6400.0}, '5': {'precision': 0.760898282694848, 'recall': 0.9, 'f1-score': 0.8246241947029349, 'support': 6400.0}, '6': {'precision': 0.9897706137631742, 'recall': 0.9978125, 'f1-score': 0.9937752878929349, 'support': 6400.0}, '7': {'precision': 0.5691368959748057, 'recall': 0.90359375, 'f1-score': 0.6983877785157901, 'support': 6400.0}, '8': {'precision': 0.9998436522826767, 'recall': 0.99921875, 'f1-score': 0.9995311034698343, 'support': 6400.0}, '9': {'precision': 0.5366459627329192, 'recall': 0.3375, 'f1-score': 0.41438848920863314, 'support': 6400.0},
  38. '10': {'precision': 0.5114308018289283, 'recall': 0.7165625, 'f1-score': 0.5968634086028504, 'support': 6400.0},
  39. 'accuracy': 0.711307661680223, 'macro avg': {'precision': 0.7129238963911154, 'recall': 0.7116056616928296, 'f1-score': 0.6840194182975478, 'support': 70324.0}, 'weighted avg': {'precision': 0.7126261386003886, 'recall': 0.711307661680223, 'f1-score': 0.6836910146192221, 'support': 70324.0}}
  40. 整体准确率: 0.71
  41. ==============================
  42. dnn_classification
  43. Macro-average Precision: 0.7185714285714287
  44. Macro-average Recall: 0.7121428571428573
  46. Process finished with exit code 0
复制代码
混淆矩阵

所有代码下载

  1. https://docs.qq.com/sheet/DUEdqZ2lmbmR6UVdU?tab=BB08J2
复制代码


免责声明：如果侵犯了您的权益，请联系站长，我们会及时删除侵权内容，谢谢合作！更多信息从访问主页：qidao123.com:ToB企服之家，中国第一个企服评测及商务社交产业平台。

本帖子中包含更多资源

您需要 登录 才可以下载或查看，没有账号？立即注册

x
回复

使用道具 举报

0 个回复

正序浏览
返回列表

快速回复

高级模式
B Color Image Link Quote Code Smilies
您需要登录后才可以回帖 登录 or 立即注册

本版积分规则

温锦文欧普厨电及净水器总代理

论坛元老
这个人很懒什么都没写!

楼主热帖

标签云

AI 运维 CIO 存储 服务器
微信订阅号
微信服务号
微信客服
小程序
H5
关于我们 商务合作 网站地图
©Copyright 2022-2024 杭州祥升科技有限公司 版权所有 浙ICP备20004199号
快速回复 返回顶部 返回列表