Course contentsShow
← Mathematics
Lesson 3004 of 3,12159. Statistics — Descriptive and InferentialFree lesson

Introduction to Maximum Likelihood Estimation

Understanding the likelihood function and the principle of choosing parameters that maximize observed data probability.

Introduction to Maximum Likelihood Estimation

You'll learn: How to choose parameter values that make your observed data most probable.

The Core Idea

Maximum Likelihood Estimation (MLE) is a method for estimating population parameters by asking: "Which parameter values would make the data we actually observed most likely to occur?"

Think of it like detective work. You've observed some evidence (your data), and you're asking: "What underlying conditions (parameters) would make this evidence most believable?" If you flip a coin 10 times and get 7 heads, MLE asks: "What value of p (probability of heads) makes these 7 heads most likely?"

The Likelihood Function

The likelihood function L(θ) measures how probable your observed data is under different parameter values θ. It's built from the probability distribution of your data, but viewed as a function of the parameters rather than the data.

For a sample x₁, x₂, ..., xₙ from a distribution f(x; θ):

  • L(θ) = f(x₁; θ) × f(x₂; θ) × ... × f(xₙ; θ)

Unlike the Method of Moments (which matches sample moments to population moments), MLE directly maximizes the probability of observing your specific sample.

The Principle

The maximum likelihood estimate θ̂ is the parameter value that maximizes L(θ). We choose the parameters that make our observed data "most at home" in the distribution.

For example, if you sample n values with mean x̄ from a normal distribution, MLE will estimate μ = x̄—because that value of μ maximizes the likelihood of seeing exactly those data points.

Key Takeaway: Maximum Likelihood Estimation finds parameter values that maximize the probability of the observed data under the assumed model.