Isolation Forest dengan Exploratory Data Analysis pada Anomaly Detection untuk Data Transaksi

Abstrak

Managing value of data is one of the key aspects of presenting analysis for decision making support in various cases. One of such method is by managing detecting anomaly in the data. This research focuses on implementing Isolation Forest result of anomaly detection. This method is used on transaction dataset from Kaggle with about more than 500.000 records. The result this research shows that Isolation Forest used in the dataset have 0.899 in accuracy, 0.00649 in precision, 0.504 in recall, and 0.013 in F1 score.