* Evaluate the difference between data transformation techniques
* Is PCA better than Kernel PCA? * Is silhouette score best metric do use, try different evaluation metrics and comment on the result * Try all unsupervised algorithms that you studied * Compare between EM and DBSCAN and isolated random forest as anomaly detection algorithm * Justify all your chooses and comment on every result * Show how result of T-SNE differs with every choose you made
- Log Transformation - Clipping Method - Scaling Methods
- PCA vs Kernel PCA
- Kmeans vs Hierarchical clustering
- EM vs DBscan vs Isolated RF - anomaly detection
- Try different evaluation metrics + T-SNE
