K-Means Clustering Kya Hai? Shuruaati Guide (Hindi)

Updated · By Viviai

K-means clustering ek unsupervised machine learning algorithm hai jo bina label wale data points ko K groups (clusters) mein baant-ta hai, jahan har group ka ek center (centroid) hota hai aur har data point apne sabse najdeeki centroid wale group mein jaata hai.

Clustering kya hai

Clustering mein machine khud data mein patterns dhoondhti hai aur milti-julti cheezon ko ek group mein rakh deti hai, bina kisi pehle se diye label ke. K-means is kaam ke liye sabse zyada istemaal hone wale algorithms mein se ek hai.

K-means kaam kaise karta hai

'K' ka matlab hai groups ki sankhya, jo aap pehle se tay karte hain. Algorithm ke steps is tarah hain:

  1. K centroids ko graph par randomly rakha jaata hai.
  2. Har data point ko uske sabse najdeeki centroid ke group mein daala jaata hai.
  3. Har group ke points ka mean nikaal kar centroid ko naye sthaan par shift kiya jaata hai.
  4. Steps 2 aur 3 tab tak dohraye jaate hain jab tak points ka group badalna band na ho jaaye.

Naye centroid ka formula hai:

[ c_k = \frac{1}{n_k}\sum_{i=1}^{n_k} x_i ]

Centroid kya hota hai

Centroid (केंद्रक) ek group ka representational center hota hai. K-means mein yeh usually us group ke saare points ka mean hota hai. Har point us centroid ke paas jaata hai jisse uski doori sabse kam ho.

Sahi K kaise chunein

K-means mein K ki value pehle deni padti hai. Ek aam tareeka elbow method hai: alag-alag K values par inertia, yaani within-cluster sum of squares, ko plot kiya jaata hai. Jahan curve ek "elbow" jaisa mod leta hai, wahi K ka ek reasonable choice maana jaata hai.

Kahan istemaal hota hai

Istemaal Example
Customer segmentation Graahakon ko umar aur kharch ke aadhar par group karna
Image compression Har pixel ko ek representative color se map karna
Document clustering Milti-julti cheezon wale documents ko group karna
Anomaly ki taraf ishara Kisi bhi group ke theek se fit na hone wale points

Kya-kya seemaayein hain

K-means ko K pehle se bataana padta hai, aur yeh shuruaati centroids par sensitive hota hai. Outliers mean ko kheench sakte hain, isliye jab data mein bahut zyada outliers hon ya clusters alag size ke hon to results ki jaanch zaroori hai.

FAQ

K-means supervised hai ya unsupervised?

Yeh unsupervised algorithm hai, kyunki yeh bina label wale data par groups banata hai.

Centroid kya hota hai?

Ek cluster ka center, jo aam taur par us group ke points ka mean hota hai.

Elbow method kya karta hai?

Yeh inertia curve par mod (elbow) dhoondh kar K ki reasonable value suggest karta hai.

K-means outliers se kyun prabhavit hota hai?

Kyunki centroid mean se banta hai, aur ek outlier mean ko apni taraf kheench sakta hai.

Mukhya baatein

Aage padhiye unsupervised learning kya hai aur machine learning kya hai.

Sources

  1. https://aigyani.in/blog/k-means-clustering-explained
  2. https://www.dswallah.com/learn/data-science/clustering/
  3. https://primelearnpoint.com/machine-learning/k-means-clustering/

Learn by doing

Read a guide, then practice with bite-sized AI lessons you can take offline.

Start learning free