Initialization of centroids, K-means++

แชร์
ฝัง
  • เผยแพร่เมื่อ 9 ก.พ. 2025
  • "(1) Initialization of centroids
    (2) Probabilistic initialization of centroids in K-means++"

ความคิดเห็น • 8

  • @SiddhantSethi02
    @SiddhantSethi02 ปีที่แล้ว +16

    I feel the explanation of the theory was very nice and precise, however, it was very difficult to get an intuition. My suggestion would be to supplement the theory with visuals.
    I really appreciate the hard work that put in bringing forward the theory and I have always been a fan of IIT content. :D :D

    • @pranshujaiswal2396
      @pranshujaiswal2396 ปีที่แล้ว +2

      Yes. The only thing lacking in these lectures is a proper visual example. The teaching quality is absolutely perfect otherwise.

    • @lakshyajain113
      @lakshyajain113 ปีที่แล้ว +1

      Definitely giving 1-2 examples in between would we a great hep

    • @hs-iitm
      @hs-iitm 8 หลายเดือนก่อน +1

      @@lakshyajain113 He gave , but only by mouth, not visually. I suggest make notes, convert his words to sentences on paper. It helps.

  • @abhinavprakashgupta89
    @abhinavprakashgupta89 ปีที่แล้ว +1

    How about we each co-ordinate of the data point and then take a sum the normalized co-ordinates and make a single value . Then this value we can look at nth percentile intervals where the step size between two nth intervals will depend upon 100 / (count of clusters needed)

  • @vatg2001
    @vatg2001 2 ปีที่แล้ว +1

    Hello 🙋‍♂️