With your last example, if you want to compute 4 clusters instead of 3, would you compare the SSE of all three clusters again and split the one with the highest? Or only the SSE of the two that are splitted at last?
Why did you rename c2 to c1' and c1'', and then rename c1 to c2 when going from step 2 to 3? Shouldn't it just be c1 becomes c1' and c1'' and c2 stays as c2?
![I.N "HALLUCINATION" | [Stray Kids : SKZ-PLAYER]](http://i.ytimg.com/vi/n5B5q1Hwt_U/mqdefault.jpg)
For notes👉 github.com/ranjiGT/ML-latex-amendments
bhut bdiya,concepts bhut acche se btaaya hai
Excellent 👍
You are technically sound and logically consistent
Thanks for uploading
Very Helpful video 👍👌
technically sound lmao
really helpful
Sir please give the mathematical intusion behind it and how it work with real dataset
I think in the formula for SSE you meant from i=1 to the number of points in the cluster instead of from i=1 to k
Nice video, thank you
With your last example, if you want to compute 4 clusters instead of 3, would you compare the SSE of all three clusters again and split the one with the highest? Or only the SSE of the two that are splitted at last?
Compare SSE of all three then split
May I know what is the formula of SSE
Why did you rename c2 to c1' and c1'', and then rename c1 to c2 when going from step 2 to 3? Shouldn't it just be c1 becomes c1' and c1'' and c2 stays as c2?
how it is guaranteed to achieve global minimum point
It depends on how many data points you have and also the number of clusters you desire to learn from.
Can't see the board. You are standing in front of it Sir, lol