Clustering

Which of the following distance measures is most sensitive to outliers?

• Euclidean

• Manhattan/ City Block

• Correlation Based Similiarity

Answer explanation

Outliers, which are data points significantly different from the rest, can have a substantial impact on the squared differences. Since Euclidean distance considers the squared differences, outliers contribute disproportionately to the overall distance. This sensitivity can lead to distorted measurements, as outliers can dominate the calculation and affect the overall similarity or dissimilarity between data points.

Which clustering progressively combining the two nearest clusters until there is just one cluster left at the end

• K-Means Algorithm

• Agglomerative Clustering

• Divisive Clustering

Answer explanation

Agglomerative methods begin with n clusters and sequentially merge similar clusters until a single cluster is obtained.

Divisive methods work in the opposite direction, starting with one cluster that includes all records.

K-Means: Using a pre-specified number of clusters, the method assigns records to each cluster.

Can decision trees be used for performing clustering?

• Yes

• No

Answer explanation

There is a variant of decision trees called "Random Forests" that can indirectly be used for clustering.

In a Random Forest, each tree is constructed using a random subset of the data and a random subset of features. While each tree individually performs a classification task, you can use the forest to identify clusters in the data based on the similarity of instances in the feature space.

What is a Dendrogram?

• 
  A type of hierarchical clustering

• 
  A type of bar chart diagram to visualize k-means clusters

• A tree diagram used to illustrate the arrangement of clusters in hierarchical clustering

What is true about K-Mean Clustering?

1. 1. K-means is extremely sensitive to cluster center initializations

2. 2. Bad initialization can lead to Poor convergence speed

3. 3. Bad initialization can lead to bad overall clustering

• 1

• all of the above

• 2 and 3

• 3

What should be the best choice for the number of clusters based on the following results?

• 14

• 6

• 4

• 10

Answer explanation

Based on the above results, the best choice of the number of clusters using the elbow method is 6. Look for an "elbow" in the plot, where the rate of decrease sharply changes. The point where this occurs is often considered the optimal k.

What is the primary goal of clustering in machine learning?

• Discovering inherent groupings or patterns in the data

• Predicting a continuous variable

• Classifying data points into predefined categories

Answer explanation

Goal: segment the data into a set of homogeneous clusters of records for the purpose of generating insight

In k-means clustering, what does "k" represent?

• Number of clusters

• Number of iterations

• Number of dimensions