Data Analysis: Clustering and Classification (Lec. 1, part 3)

[@KKdudle]

You are the best, all is simple and easy to understand. Thank you for your videos and your time you put in to them.

[@GarrettGaudini]

This is awesome! so helpful to someone with a 20 year old MSE degree jumping back into ML.

[@madisonboudreaux4470]

I know im not a biology major, but this is so helping in my computer science Data mining Elective.

[@heyyou827]

This is so helpful. Not a biology major but this is helping with my work which involves data :)

[@troooooper100]

Problem, albeit edge case, with K-Nearest Algorithm is that order of new data affects out come. You could have same data but in what order you add it affects what end result looks like. at 7:04 suppose there are 5 new data points in middle stacked vertically and very close to each other, so much that those 5 are always nearest neighbor to each other. now if you start from top point in this stack, then it'd be circle, next point will also become circle since p1 was circle and other two are 1 box and 1 circle... now if you had started form bottom you'd get a box first, which would influence next 4 ... domino effect. this is hard edge case, but to me it's also demonstrate inherent flaw.

[@baharehbahrami3538]

Thanks for your wonderful lecture so easy to undrestand and the way of your presenting is uniqe

[@jayalalkj1576]

Thanks-good job.

[@zapshadab3450]

can you just give information about which the very efficient algorithm belong these..

[@shobhamourya8396]

If there are n labels, k is n+1