Research Article
An Optimized Computational Framework for Isolation Forest
Table 3
The AUC results of the five methods on eight datasets. Values marked in bold are the best results.
| Date sets | LOF | iForest | SCiForest | OIT | OIF |
| Breast | 0.4940.132 | 0.4960.149 | 0.5130.128 | 0.5580.152 | 0.560±0.126 | Ionosphere | 0.8930.093 | 0.8550.083 | 0.916±0.071 | 0.8430.074 | 0.8740.069 | Pima-diabetes | 0.5870.051 | 0.6790.053 | 0.6310.036 | 0.6490.058 | 0.734±0.055 | Breast-diagnostic | 0.5250.086 | 0.7940.069 | 0.8530.053 | 0.6710.109 | 0.883±0.050 | CreditCardDefault | 0.4990.014 | 0.5920.017 | 0.5870.014 | 0.6190.017 | 0.653±0.013 | PenDigits | 0.6500.193 | 0.7960.088 | 0.8020.100 | 0.8650.128 | 0.938±0.035 | UnionPay | 0.4680.056 | 0.6450.052 | 0.4720.066 | 0.6010.060 | 0.716±0.057 | Kddcup99 | 0.6160.110 | 0.989±0.006 | 0.9870.009 | 0.9620.039 | 0.9860.010 |
|
|