Risk_Level_Classification is a dataset for classifying transaction risk levels, updated by Jack Ward. The target variable 'anomaly' is treated as a nominal variable with three categories: low risk, moderate risk, and high risk. The dataset is licensed under CC-BY-4.0 and is hosted on OpenML.
Use Cases
- Training a classifier to predict transaction risk levels based on the 'anomaly' target variable.
- Benchmarking anomaly detection algorithms on a dataset with three distinct risk categories.
- Developing fraud detection systems that categorize transactions by risk severity.
Strengths
- The target variable 'anomaly' is explicitly defined with three nominal categories: low, moderate, and high risk.
- The dataset is an updated version, suggesting potential improvements over a prior iteration.
- It is published under the permissive CC-BY-4.0 license.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment for large-scale training.
- Last update date is unknown; freshness unverified.
Provenance
- Source
- Jack Ward via OpenML
- Collection Method
- null
- Time Range
- null
- Freshness
- null
- Geography
- null