CPU Anomaly Detection System — ML-2

Overview

Real-time CPU monitoring system that detects anomalies using Statistical and ML methods.

Dataset

• cpu_utilization_asg_misconfiguration.csv
• 18,050 readings | May 2014 → July 2014

Methods Used

• Statistical Detection (Mean + 2σ threshold)
• Isolation Forest (Scikit-learn)
• PyOD Anomaly Detection

Results

• Total Anomalies: 1,450 (8.03%)
• Critical: 826 | Warning: 624

Tech Stack

Python, Pandas, NumPy, Scikit-learn, PyOD, Matplotlib

How to Run

1. Open notebook in Google Colab
2. Upload dataset from Data/ folder
3. Run all cells step by step
4. Check Output/ for results