Normal Distribution Explained

in standart deviation , we know 68.8% data lies in 1

standard deviation , 95% in 2 std . is there any proof that 1
std contain that percentage data and so on

Yes! The 68-95-99.7 Rule (Empirical Rule) comes from the properties of the normal
distribution and can be proven using integrals of the normal probability density function
(PDF).

1️⃣ The Standard Normal Distribution

A normal distribution is given by:

1 −
(x−μ)2
f (x) = e 2σ2 ​

σ 2π
​

where:

μ is the mean
σ is the standard deviation
The area under the entire curve is 1 (since it's a probability distribution).

2️⃣ Proof Using Integration

To find the probability that data lies within 1, 2, or 3 standard deviations, we integrate the
standard normal distribution:

b
1 −z 2 /2
P (a ≤ Z ≤ b) = ∫ e dz
2π
​ ​

a ​

X−μ
where Z is the standard normal variable (Z = σ ).

📌 Computing the probabilities:

​

1. For ±1σ (68.27%)

1/6
 1
1 −z 2 /2
P (−1 ≤ Z ≤ 1) = ∫ e dz ≈ 0.6827
2π
​ ​

−1 ​

2. For ±2σ (95.45%)

2
1 −z 2 /2
P (−2 ≤ Z ≤ 2) = ∫ e dz ≈ 0.9545
2π
​ ​

−2 ​

3. For ±3σ (99.73%)

3
1 −z 2 /2
P (−3 ≤ Z ≤ 3) = ∫ e dz ≈ 0.9973
2π
​ ​

−3 ​

These values match the Empirical Rule.

3️⃣ Alternative Proof Using the Cumulative Distribution Function

(CDF)
Instead of integrating manually, we can use the CDF of the normal distribution, Φ(z), which is
precomputed:

P (−k ≤ Z ≤ k) = Φ(k) − Φ(−k)

Using standard normal tables:

Φ(1) − Φ(−1) = 0.8413 − 0.1587 = 0.6827 (~68%)

Φ(2) − Φ(−2) = 0.9772 − 0.0228 = 0.9545 (~95%)
Φ(3) − Φ(−3) = 0.99865 − 0.00135 = 0.9973 (~99.7%)

4️⃣ Real-Life Example & Verification

Would you like me to generate a dataset and plot how real data follows this rule? I can
simulate normally distributed data and verify the percentages within 1, 2, and 3 standard
deviations. 🚀
yess

2/6
python

import numpy as np Always show details

import matplotlib.pyplot as plt

import seaborn as sns

# Generate a large sample of normally distributed data (mean=100, std=15)

mean, std, size = 100, 15, 10000
data = np.random.normal(mean, std, size)

# Compute the standard deviation ranges

one_std_range = (mean - std, mean + std)
two_std_range = (mean - 2*std, mean + 2*std)
three_std_range = (mean - 3*std, mean + 3*std)

# Count percentages within each range

within_one_std = np.mean((data >= one_std_range[0]) & (data <= one_std_range[1])) *
100
within_two_std = np.mean((data >= two_std_range[0]) & (data <= two_std_range[1])) *
100
within_three_std = np.mean((data >= three_std_range[0]) & (data <=
three_std_range[1])) * 100

# Plot histogram with KDE

plt.figure(figsize=(10, 5))
sns.histplot(data, bins=50, kde=True, color='blue', edgecolor='black', alpha=0.6)

# Highlight standard deviation regions

plt.axvline(one_std_range[0], color='red', linestyle='dashed', label=f'±1σ
({within_one_std:.2f}%)')
plt.axvline(one_std_range[1], color='red', linestyle='dashed')

plt.axvline(two_std_range[0], color='orange', linestyle='dashed', label=f'±2σ

({within_two_std:.2f}%)')
plt.axvline(two_std_range[1], color='orange', linestyle='dashed')

plt.axvline(three_std_range[0], color='green', linestyle='dashed', label=f'±3σ

({within_three_std:.2f}%)')
plt.axvline(three_std_range[1], color='green', linestyle='dashed')

# Labels and legend

plt.title("Normal Distribution & Standard Deviation Ranges")
plt.xlabel("Value")

3/6
plt.ylabel("Frequency")
plt.legend()
plt.show()

# Return computed percentages

within_one_std, within_two_std, within_three_std

python

# Re-import necessary libraries since execution state was reset Always show details

import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns

# Generate a large sample of normally distributed data (mean=100, std=15)

mean, std, size = 100, 15, 10000
data = np.random.normal(mean, std, size)

# Compute the standard deviation ranges

one_std_range = (mean - std, mean + std)
two_std_range = (mean - 2*std, mean + 2*std)
three_std_range = (mean - 3*std, mean + 3*std)

# Count percentages within each range

within_one_std = np.mean((data >= one_std_range[0]) & (data <= one_std_range[1])) *
100
within_two_std = np.mean((data >= two_std_range[0]) & (data <= two_std_range[1])) *
100
within_three_std = np.mean((data >= three_std_range[0]) & (data <=
three_std_range[1])) * 100

# Plot histogram with KDE

plt.figure(figsize=(10, 5))
sns.histplot(data, bins=50, kde=True, color='blue', edgecolor='black', alpha=0.6)

# Highlight standard deviation regions

plt.axvline(one_std_range[0], color='red', linestyle='dashed', label=f'±1σ
({within_one_std:.2f}%)')
plt.axvline(one_std_range[1], color='red', linestyle='dashed')

plt.axvline(two_std_range[0], color='orange', linestyle='dashed', label=f'±2σ

4/6
 ({within_two_std:.2f}%)')
plt.axvline(two_std_range[1], color='orange', linestyle='dashed')

plt.axvline(three_std_range[0], color='green', linestyle='dashed', label=f'±3σ

({within_three_std:.2f}%)')
plt.axvline(three_std_range[1], color='green', linestyle='dashed')

# Labels and legend

plt.title("Normal Distribution & Standard Deviation Ranges")
plt.xlabel("Value")
plt.ylabel("Frequency")
plt.legend()
plt.show()

# Return computed percentages

within_one_std, within_two_std, within_three_std

Result

(67.86999999999999, 95.56, 99.76)

Here’s the histogram of the normally distributed data along with the standard deviation
ranges. The computed percentages of data within each range are:

1 standard deviation (±1σ): ≈ 67.87% (Expected: 68.27%)

2 standard deviations (±2σ): ≈ 95.56% (Expected: 95.45%)

5/6
 3 standard deviations (±3σ): ≈ 99.76% (Expected: 99.73%)

These values closely match the theoretical Empirical Rule, confirming that a normal
distribution follows these probabilities in real-world data. 🚀​

6/6