Understand Cosine Similarity | 2 Minute Tutorial

This is a quick introduction to cosine similarity - one of the most important similarity measures in machine learning!

Cosine similarity meaning, formula and example!

If you like this video, hit a like and subscribe!

Paper icon used in a video:

[einstein-munachi]

Such a simplistic yet on-point explanation.... cheers mate!

[maxreed]

Brilliantly concise explanation. Have a like. :)

[minlingg]

useful video and clear explanation! thank you and keep up the good work!

[AV_YOUTUBE_X]

I would say, pre-weight each vector component somehow (such as via examples of performance of the similarity metric you desire, and then calculating what each component weight should be in order to make the examples perform like expected). This might help, because certain components are going to have more important meaning, whereas certain components are going to be more irrelevant. Then, *after* you pre-weight the components of your inputs (premultiply each input component by each weight component), *then* measure their cosine similarity. That's probably the best similarity metric you can get, I would expect.

[aleefbilal]

bro looks 18 and 30 at the same time.
anyways, great and quick explanation. Thanks.

[desecrator]

Thanks for the quick and crisp explanation

[einstein-munachi]

Is the magnitude of B ~= 1.732 or ~= **2.236**

[galacticimaginarium]

Indeed pretty amazing explaination, helped me a lot! Thanks.

[soberian]

Great and efficient explanations, you deserve million views, this help me alot. Can you explain about K-Nearest Neighbors? I would love to watch your explanation.

[simo_woman]

Great having it explained in short! Thanks and waiting for the next one!

[_cup]

Thank you for the straightforward examples. On point!

[omkarrajmane]

Wow, wan't expecting much as it was such a short video. But this was the most valuable video that was so concise and made me understand the concept much better! Thanks a lot !

[exoticcoder]

Very useful to have a quick recall on the calculation part 👍

[Wilhuf]

Oh I thought cosine similarity ranged only between 0.0 and 1.0.

[EjazAhmed-pftz]

unbelievable
i have tommorrow exam and still writing this comment which i do not usually
thank you so much for such simple explanation and explaining 2 hour lecture in just 2 minutes thank you once again

[nikola]

Can you explain why 3 vectors? Because there are three sentences? Then you got v1, v2, v3 and v1, v2 and v1, v3, why there is no v2, v3? Is it always first and then all others?

[toastrecon]

I was listening to a talk the other day, and someone mentioned that cosine similarity might be replaced eventually by something called "learned representation"? I may have it wrong, but have been struggling to find any info on it. Have you heard of that?