Review Mode Set 1 – AI-900 – Question 13

[Dave Crewe]

Hi, seeking some clarity on the third question.

“Sorting songs into playlists based on their genres is an example of clustering” – The correct answer in the quiz is ‘Yes’, but I am confused as to why this isn’t ‘No’ as this looks like classification, not clustering.

Further down In the answer we see the following:

“<strong style=”font-family: inherit; font-size: inherit;”>Clustering groups similar data points together based on feature similarities, enabling data exploration and customer segmentation without predefined labels.”

But song genres ARE predefined labels, so the answer should be ‘No’ shouldn’t it?

[Nikee-TutorialsDojo]

Hi Dave,

Great question! This is a common point of confusion between classification and clustering.

Clustering is an unsupervised learning technique where data points are grouped based on similarities in their features, without using predefined labels. For example, if you had only song attributes (tempo, rhythm, energy, etc.) and let an algorithm automatically group songs into “clusters,” that would be clustering. The groups might later resemble genres, but they wouldn’t start with genre labels.

Classification, on the other hand, is a supervised learning task where the model is trained with predefined labels (such as “rock,” “jazz,” “pop”) and learns to assign new data points into one of those categories.

So why does the quiz say “Yes” for clustering? The intent is to test your understanding that clustering = grouping data by feature similarity. Here, the feature is the genre of the songs. Therefore, similar songs (same or similar genre) are grouped, which is the essence of clustering.

I hope this clears up your confusion. If you need further assistance, please don’t hesitate to contact us.

Regards,

Nikee @ Tutorials Dojo

[Dave Crewe]

Thanks for the clear and detailed response, much appreciated!

[Nikee-TutorialsDojo]

Happy to help, Dave!

Cheers,

Nikee