k means++ initial center selection visualization

Поділитися
Вставка
  • Опубліковано 28 сер 2024
  • Visualization shows how K means++ selects the initial points (for k=4)
    For more info check
    Updated : • K-Means++ Centroid Ini...
    en.wikipedia.o...

КОМЕНТАРІ • 16

  • @nevilmehta2281
    @nevilmehta2281 9 місяців тому +9

    1 minute of animation taught me more than half an hour of a lecture! Thanks a lot!

  • @100_IQ_EQ
    @100_IQ_EQ 2 місяці тому

    Thank you for the video. Explains everything clearly except the choice of third point and how many centers are chosen.

    • @crazyland7507
      @crazyland7507  2 місяці тому

      Number of centers is a parameter given as input to this algorithm. I have selected 4.

  •  Рік тому +6

    That was perfectly illustrated example. Very much appreciated the effort you've put on it. Thanks!

  • @ponselvanarumugasamy3640
    @ponselvanarumugasamy3640 Рік тому

    Awesomely illustrated
    Some people are making long videos still couldn't able to explain it clearly
    Thank you so much

  • @aarohizade5878
    @aarohizade5878 6 місяців тому

    elegant in its simplicity. thanks a million

  • @markocska94
    @markocska94 Рік тому

    Wow! So much more intuitive when I see it. Thank you!

  • @matinfazel8240
    @matinfazel8240 Рік тому

    It was Awsome and well-illustrated.

  • @iamacoder8331
    @iamacoder8331 Рік тому

    Oh man, that is crazy, thanks

  • @ArjunSK
    @ArjunSK 11 місяців тому

    Great animation!

  • @EjazAhmed-pf5tz
    @EjazAhmed-pf5tz Рік тому

    wow amazing

  • @thaimeuu
    @thaimeuu 9 місяців тому

    thanks bro

  • @mustafaerogluu
    @mustafaerogluu 9 місяців тому

    After choosing the second center, how do we choose the third center?

    • @crazyland7507
      @crazyland7507  9 місяців тому +2

      Once we select the second point, each point is taken and distance from both of the centers is calculated.
      As we see in the animation each point has 2 distance associated.
      Whichever distance is smaller the point is considered to be a part of that cluster, you can see at 00:36 all points are now associated with one of the 2 clusters.
      Now we take the point that is furthers from the center or you can say we take the longest line from 00:36 , this point will become the third center.

    • @100_IQ_EQ
      @100_IQ_EQ 2 місяці тому

      ​@@crazyland7507This was unclear to me as well. But, I think the way is to find the farthest point from the centre of the cluster 1 within cluster 1. Call it F1. Centre of cluster 1 is C1. Repeat the same with cluster 2. Centre of cluster 2 is C2. Farthest point is F2.
      Compare the distances, F1-C1 and F2-C2. Let's call the distances D1 and D2 respectively.
      If D1 is greater, F1 becomes the third centre. And if D2 is greater, F2 becomes the third centre.
      Then, form three clusters. Again find Farthest points. Find distances. Choose 4th centre. This process continues until a stopping criteria is met. Suppose the Distances between farthest points and the cluster centers within the cluster are below a certain threshold, algorithm stops. (Let's call these distances farthest distances.) Or the farthest distances within the clusters is a fraction of shortest of distances between cluster centers etc.