CPSC 330 Lecture 16: DBSCAN, Hierarchical Clustering

Andrew Roth (Slides adapted from Varada Kolhatkar and Firas Moosvi)

Announcements

• MT2 next week! See Piazza.
• Final exam dates and registration posted. See Piazza.
• HW6 is posted

16.1 Select all of the following statements which are True (iClicker)

iClicker join link: https://join.iclicker.com/HTRZ

• 1. Similar to K-nearest neighbours, K-Means is a non parametric model.
• 2. The meaning of K in K-nearest neighbours and K-Means clustering is very similar.
• 3. Scaling of input features is crucial in clustering.
     
• 4. In clustering, it’s almost always a good idea to find equal-sized clusters.

Class demo

Class notebook

16.2 Select all of the following statements which are True (iClicker)

iClicker join link: https://join.iclicker.com/HTRZ

• 1. With tiny epsilon (eps in sklearn) and min samples=1 (min_samples=1 in sklearn) we are likely to end up with each point in its own cluster.
• 2. With a smaller value of eps and larger number for min_samples we are likely to end up with one big cluster.
• 3. K-Means is more susceptible to outliers compared to DBSCAN.
     
• 4. In DBSCAN to be part of a cluster, each point must have at least min_samples neighbours in a given radius (including itself).

Class notebook

16.3 Select all of the following statements which are True

iClicker join link: https://join.iclicker.com/HTRZ

• 1. In hierarchical clustering we do not have to worry about initialization.
• 2. Hierarchical clustering can only be applied to smaller datasets because dendrograms are hard to visualize for large datasets.
• 3. In all the clustering methods we have seen (K-Means, DBSCAN, hierarchical clustering), there is a way to decide the granularity of clustering (i.e., how many clusters to pick).
• 4. To get robust clustering we can naively ensemble cluster labels (e.g., pick the most popular label) produced by different clustering methods.
• 5. If you have a high Silhouette score and very clean and robust clusters, it means that the algorithm has captured the semantic meaning in the data of our interest.