Clustering

Question 1

What is the disadvantage of KMeans?

Answer 1

1. No. of clusters is not specified.
2. Even if we are using elbow method sometimes the elbow is not clear from the plot
3. Sensitive to outliers (As it is a distance based algorithm)
4. Perform poor if data is not spherical

Question 2

Full form of DBSCAN

Answer 2

Density Based Spatial Clustering Application with Noise

Question 3

Epsilon

Answer 3

Radius of circle

Question 4

MinPts

Answer 4

Minimum no. of points in circle for it be called as a core point.

Question 5

Core point

Answer 5

Point which has more or equal MinPts in Epsilon.

Question 6

Border point

Answer 6

Point which has less than MinPts in Epsilon but at least one core point.

Question 7

Noise point

Answer 7

Points other than core and border

Question 8

Advantages and disadvantages of DBSCAN

Answer 8

Advantages:
1. Robust to outliers
2. No need to specify no. of clusters
3. Can find arbitrary shaped clusters

Disadvantages:
1. Difficulty with varying density clusters
2. Cannot predict cluster for a new point

Question 9

Types of Hierarchical Clustering

Answer 9

1. Agglomerative Clustering (Bottom-up)
2. Divisive Clustering (Top-down)

Question 10

Proximity matrix

Answer 10

A square matrix that stores the distances between each pair of data points

Question 11

Types of agglomerative clustering

Answer 11

1. Min (Single link)
2. Max (Complete link)
3. Average
4. Ward

Question 12

Single Link

Answer 12

Finds the minimum distance between any two points of the cluster.
The distance is found out by taking the combination of all possible distances.

Should not be used in presence of outliers.

Question 13

Complete Link

Answer 13

Finds the maximum distance between any two points of the cluster.
The distance is found out by taking the combination of all possible distances.

Robust to outliers
However, in case of different sized cluster, the bigger cluster may break in smaller sections.

Question 14

Group Average

Answer 14

Finds the average distance between every two points of the clusters.

Trade off between Single and complete link.

Question 15

Ward’s Method

Answer 15

The similarity of two clusters is based on the increase in squared error when two clusters are merged.

Question 16

How to find the ideal number of clustering in agglomerative clustering?

Answer 16

In dendrogram plot, where the vertical line (without any horizontal cut) is the longest.

Question 17

Advantages and disadvantages of Agglomerative

Answer 17

Advantages:
1. Widely applicable
2. Dendrogram

Disadvantages:
1. Depends on type of linkage & choice of dissimilarity.
2. Computationally expensive for larger datasets, due to distance calculation
3. Even outliers are forced to a cluster

Question 18

How is PCA different from clustering?

Answer 18

PCA is a low-dimensional representation of data, whereas Clustering is categorizing the data into homogeneous subgroups.

Question 19

Dendrogram

Answer 19

Tree like visual representation of the observation

Question 20

Can we cluster features on the basis of observations to discover subgroups among of the feature

Answer 20

Yes

Question 21

What is the criteria to determine good cluster in K-Means?

Answer 21

WCSS - With in cluster sum of squared distance (Euclidean)

Question 22

Centroid linkage

Answer 22

Dissimilarity between centroid of two clusters.

Can result in inversion

Question 23

Choice of dissimilarity measure

Answer 23

1. Euclidean distance
2. Correlation based distance - Two observations are similar if they are highly correlated
   It focusses on shape of observation rather than their magnitudes.

Question 24

Should the observations first be standardized in any way before clustering?

Answer 24

Depends on the problem we are trying to solve.

Question 25

For a new point, how different clustering algorithm works?

Answer 25

K-Means: Yes - Using centroid of initial clusters
DBSCAN: No, need to re-run the algorithm
Agglomerative Clustering: No, need to re-run the algorithm