Cluster Test - Machine Learning

Question 1

One of the main differences between Segmentation and Classification Analysis is that Segmentation goal is to find groups

Answer

true

Question 2

Cluster is the only technique you can use to Segment

Answer

true

Question 3

Standardization of variables is a common task before clustering a multivariate dataset

Answer

true

Question 4

Two-Step (Interactive) Cluster output can be conditioned by statistical hypothesis about the probability distributions of clustering fields

Answer

true

Question 5

K-Means cluster HELPS to determine the “optimum” number of clusters

Answer

true

Question 6

K-means’ results depend on initial cluster’s centers

Answer

true

Question 7

K-Means is normally able to correctly find clusters of different sizes

Answer

true

Question 8

The most important validation criterion in Segmentation is Silhouette

Answer

true

Question 9

Profiling variables are those entered into the algorithm to find groups, and then used to describe group profiles

Answer

true

Question 10

If we are looking for a precise solution, there should be no correlation whatsoever between explanatory variables

Answer

true

Question 11

If two clusters share the same label, the segmentation result is not optimum

Answer

true

Question 12

We always need a target variable/field when running a cluster analysis

Answer

true

Question 13

To run K-means, we always need to select in advance the number of clusters to get

Answer

true

Question 14

We may use a gender variable coded M/F without further recoding as clustering variable in a K-means

Answer

true

Question 15

Silhouette measure from a Cluster output ranges from 0 to 1

Answer

true

Question 16

If we obtain several small clusters (less than 5% of total sample size), it is normally recommendable to merge them

Answer

true

Question 17

It is possible to obtain the cluster membership in Dataiku DSS

Answer

true

Create a free account to finish this quiz