Course contentsShow
← Machine Learning and Deep Learning
Lesson 67 of 3,5381. Mathematical Foundations for Machine LearningFree lesson

Normal (Gaussian) Distribution

Understand the most important continuous distribution and its properties including the 68-95-99.7 rule.

Normal (Gaussian) Distribution

You'll learn what the bell curve is, why it appears everywhere in machine learning, and how to interpret its key properties.

What Is It?

The Normal (Gaussian) Distribution is the most important continuous probability distribution in statistics and machine learning. It describes data that clusters symmetrically around a mean value, with values farther from the mean becoming progressively less likely. Think of heights in a population: most people are near average, fewer are very short or very tall.

The distribution is completely defined by two parameters:

  • Mean (μ): the center of the distribution
  • Standard deviation (σ): how spread out the values are

The Famous Bell Curve

When you plot the probability density function, you get the iconic "bell curve" — symmetric, smooth, and tailing off on both sides.

The 68-95-99.7 Rule (Empirical Rule)

This is your most practical tool for understanding normal distributions:

  • 68% of data falls within 1 standard deviation of the mean (μ ± σ)
  • 95% of data falls within 2 standard deviations (μ ± 2σ)
  • 99.7% of data falls within 3 standard deviations (μ ± 3σ)

Example Analogy

Imagine test scores with mean = 70 and standard deviation = 10:

  • 68% of students score between 60 and 80
  • 95% score between 50 and 90
  • 99.7% score between 40 and 100

Why It Matters in ML

Many real-world phenomena follow (or approximately follow) normal distributions. Neural network weight initialization often uses Gaussian distributions, and many algorithms assume normally distributed errors.

Key Takeaway: The Normal Distribution is the fundamental bell-shaped curve characterized by mean and standard deviation, with the 68-95-99.7 rule providing intuitive boundaries for where most data lies.