Research Article

An Optimized Computational Framework for Isolation Forest

Table 2

Statistics of eight datasets.

Date sets # of records# of attributesThe ratio of anomaly data

Breast19833Malignant (23.7%)
Ionosphere35133Bad (35.8%)
Pima-diabetes7688Pos (34.9%)
Breast-diagnostic56930Malignant (37.3%)
CreditCardDefault3000023yes (22.1%)
PenDigits986816yes (2%)
UnionPay307495Fraud merchants (11.5%)
Kddcup996083980Abnormal (0.4%)