Anomaly-Detection

https://benthamchang.github.io/Anomaly-Detection/Anomaly%20Detection%20Report.html

Goal

1. Detected anomolous output based on input materials via supervised learning under limited size of data.
2. Detected anomolous output based on input materials via unsupervised learning.

Analysis

1. Exploratory Data Analysis
2. Missing Value Treatment
3. Anomaly Detection
   • Linear Model
   • Isolation Forest
   • Hierarchical Clustering
   • Multivariate SPC

Conclusion

1. Identified the most critical input materials that lead to anomaly via observing linear model's p-value.
2. Detected problematic input materials and ranked them from high to low based on the predicted probability (anomolous tendency).