Fast and robust outlier detection: A granular-ball center isolation and region consistency approach

Key Points

• Outlier detection accuracy improves significantly with the proposed granular-ball method, demonstrating robustness.
• Results indicate a detection rate achieving over 90% accuracy in diverse datasets compared to traditional methods.
• The approach integrates center isolation and region consistency for effective identification of anomalies in data.
• These findings may enable better data analysis across various fields, yet further validation is necessary.