ML Knowledge
Can you suggest some ways in which the performance of a clustering algorithm can be measured when the labels are given?
Machine Learning Engineer
AT&T
IBM
VMware
Bolt
Figma
Answers
Anonymous
4 months ago
We can compare the predicted labels with the actually labels. If we are only interested in a single cluster this becomes a binary classification problem. If we are interested in all clusters, this becomes a multi class prediction problem. In either case we can calculate accuracy, precision, recall and other metrics
Interview question asked to Machine Learning Engineers interviewing at Figma, AT&T, MathWorks and others: Can you suggest some ways in which the performance of a clustering algorithm can be measured when the labels are given?.