SVM (Support Vector Machine) kya hai?

SVM (Support Vector Machine) ek powerful machine learning algorithm hai jo mainly classification aur regression problems solve karne ke liye use hota hai.

Simple Language mein samjho:

SVM ka goal hota hai data ko alag-alag categories (classes) mein divide karna — ek best boundary (line ya plane) ke through.

• Is boundary ko Hyperplane kaha jata hai.

Kaise kaam karta hai?

Maan lo tumhare paas 2 types ke data hain:

• Red points
• Blue points

SVM ek aisi line draw karta hai jo:

• Dono classes ko alag kare
• Dono classes se maximum distance banaye

Is distance ko Margin kehte hain
Aur jo points boundary ke sabse paas hote hain unhe Support Vectors kehte hain

Key Concepts:

1. Hyperplane

• Ek decision boundary (2D mein line, 3D mein plane)

2. Margin

• Hyperplane aur nearest data points ke beech ka distance
• SVM maximum margin choose karta hai (better accuracy ke liye)

3. Support Vectors

• Wo data points jo boundary ke sabse paas hote hain
• Ye model ko define karte hain

Types of SVM:

Linear SVM

• Jab data easily ek straight line se separate ho jaye

Non-Linear SVM

• Jab data complex ho
• Is case mein Kernel Trick use hota hai

Kernel Trick kya hota hai?

Agar data linearly separable nahi hai, to SVM usse higher dimension mein convert karta hai.

Popular Kernels:

• Linear
• Polynomial
• RBF (Radial Basis Function)

Real-life Example:

• Email spam detection
• Face recognition
• Text classification

Advantages:

• High accuracy
• Small dataset pe bhi achha kaam karta hai
• Overfitting ka chance kam

Disadvantages:

• Large dataset pe slow ho sakta hai
• Kernel selection tricky hota hai

Short Definition:

SVM ek supervised machine learning algorithm hai jo data ko best possible boundary (hyperplane) ke through separate karta hai, maximum margin maintain karte hue.