World Indicators Clustering
Applied K-means and hierarchical clustering methods on World Indicators dataset
The objective of this project is to implement different clustering methods to synthetic and real-world data and validate using external and internal validation techniques
Synthetic Dataset
- Applied clustering methods (K-means and hierarchical clustering) to generate clustering from 8 datasets (“Data1.csv” to “Data8.csv”)
- Evaluated the performacne of the clustering algorithm using external validation.
- Plot (2D or 3D) the data points for each dataset and color them according to the original class
- Plot (2D or 3D) the data points for each dataset and color them according to the class allocated by the clustering algorithm
Real-word data (World Indicators dataset)
- Applied K-means and hierarchical clustering methods to group similar countries together
- Used Internal validation metrics to report the cluster quality
- Conducted experiments to find appropriate numbers of clusters by using Silhouette analysis and Elbow method
Jupiter Notebook Link