Support Vector Machines (3): Kernels

[@ajitpatankar1749]

This is by far the best explanation of SVM kernel function that I have seen. It is a great balance between rigorous treatment and intuitive understanding. Thank you, Alex.

[@redobot]

Thanks for this great video, this is the first time I find a good explanation of what's the "kernel trick". Thanks for sharing your knowledge.

[@Tee_eej]

Very good set of lectures, and you are a very good presenter! Keep it up!

[@BrandonRohrer]

Great explanation Alexander. It helped some things click for me.

[@qft_k]

Great channel, great topics and very approachable. Keep posting!

[@jean-philipperoberge7718]

Thanks for this clear explanation of kernels / kernel trick!! Very good job.

[@penguinshin]

Does the kernel trick give us access to the same expressivity as a model that is learned on transformed vectors explicitely? It seems like our weights for the respective features for our polynomial transform are fixed, so, for example, if the underlying function was x1^2 + 4x1x2, we couldn't learn it since our feature is of the form x1^2 + 2x1x2 + .... etc. But if we had just concatenated all of the various components of the polynomial feature vector and ran a linear svm on that, then we would be able to learn such a function. Is this correct? or am i missing something?

[@bobby_mathews]

watching this 15 min before my presentation on SVM :)

[@pranav7471]

Wow amazing explaination

[@PierLim]

Good stuff, thank you!

[@tomgag2]

Great video

[@a.a3265]

could you please explain for me the svm by using data with a simple example

[@simprodigy]

Yes! Thank you!

[@ericaltf4]

Does anyone know what the difference is between Xi and Xj?

[@kelixoderamirez]

Permission to learn sir