Goal: Points in the same cluster are ‘close’ and points of different clusters are ’far’
⃝c -Trenn, King’s College London 2
Application: Biological Networks
Protein-Protein Interaction Networks (PPINs). See e.g., Rual et al Nature ’05 (+2500 citations)
⃝c -Trenn, King’s College London 3
Application: Recommendation Systems
Similar users will buy similar items
High dimensional data, how to find similar users? Clustering!
⃝c -Trenn, King’s College London 4
Other Applications
Unsupervised learning (e.g., PCA)
encoder-decoder neural network
outlier detection applications such as detection of credit card fraud ….
⃝c -Trenn, King’s College London 5
Clustering Algorithm (incomplete)
k-Means
k-Median
Agglomerative Clustering
‚ Single-Linkage
‚ Average-Linkage ‚ Complete-Linkage
Divisive Clustering ‚ Sparsest-Cut
⃝c -Trenn, King’s College London 6
Clustering Algorithm Comparison
Source: https://towardsdatascience.com/
⃝c -Trenn, King’s College London 7