NumPy Random Functions

NumPy Random Functions Explained (rand, randn, ranf, randint)

Random numbers are widely used in real-world applications like simulations, testing, AI models, and games.

---

🔹 Step 1: Import Libraries

import numpy as np
import random

Note: Most work here is done using np.random

---

🔹 1. rand() → Uniform Random Values

ar = np.random.rand(2,2)
print(ar)

✅ Explanation:

• Generates numbers between 0.0 to 1.0
• Values are uniformly distributed
• (2,2) → 2 rows, 2 columns

💡 Real Scenario:

Generating random probabilities → e.g., simulate user click chance in a system

⚠️ Important Points:

• rand(2) → 1D array
• rand(1,2) → 2D array
• More values = higher dimensions

---

🔹 2. randn() → Normal Distribution (Gaussian)

ar = np.random.randn(2,3)
print(ar)

✅ Explanation:

• Values are centered around 0
• Can be positive or negative
• Follows normal distribution (bell curve)

💡 Real Scenario:

Used in Machine Learning → weight initialization in neural networks

⚠️ Important:

• Most values are close to 0
• Few extreme values (like -2, +2)

---

🔹 3. ranf() → Random Float Sampling

ar = np.random.ranf(2)
print(ar)

✅ Explanation:

• Generates values between 0.0 to 1.0
• Very similar to rand()

💡 Real Scenario:

Used in simulations → random sampling of probabilities

⚠️ Important:

• Returns 1D array if single value passed
• Difference from rand() is minimal for beginners

---

🔹 4. randint() → Random Integers

ar = np.random.randint(2,10,3)
print(ar)

✅ Explanation:

• Generates integers from 2 to 9 (10 excluded)
• 3 → total numbers

💡 Real Scenario:

Generating OTPs, random IDs, dice rolls, game scores

⚠️ Important Points:

• Upper bound is always excluded
• Works only with integers

---

🔥 Key Differences (Quick View)

• rand() → uniform (0 to 1)
• randn() → normal distribution (around 0)
• ranf() → random floats (0 to 1)
• randint() → random integers

---

🚀 Pro Tips

• Use rand() for probability simulations
• Use randn() for ML models
• Use randint() for discrete values
• Set seed for same output every time:

np.random.seed(42)

👉 Useful for debugging & reproducibility

---

❌ Common Mistakes

• Confusing rand() and randn()
• Forgetting upper limit is excluded in randint()
• Assuming ranf() is very different from rand()

---

🔚 Conclusion

Random functions are essential when working with simulations, testing, and AI systems.

Mastering them helps you build real-world applications faster ⚡