Table 4.
Data sample numbers and percentage before and after oversampling CIC IDS 2017, 2018, and CIC Bell DNS 2021 datasets.
Actual | Balanced | Actual Percentage | Balanced Percentage | |||||
---|---|---|---|---|---|---|---|---|
Benign | Malicious | Benign | Malicious | Benign | Malicious | Benign | Malicious | |
CIC IDS 2017 | 2,273,097 | 557,646 | 2,273,097 | 2,230,584 | 80.3 | 19.7 | 50.47 | 49.53 |
CIC IDS 2018 | 13,484,708 | 2,748,235 | 13,484,708 | 10,992,940 | 83.07 | 16.93 | 55.09 | 44.91 |
CIC Bell DNS 2021 | 400,000 | 13,011 | 23,716 | 22,929 | 96.84 | 3.16 | 50.84 | 49.16 |