Course contentsShow
← Mathematics
Lesson 3005 of 3,12159. Statistics — Descriptive and InferentialFree lesson

Finding MLEs by Differentiation

Using calculus to maximize the log-likelihood function and solve for maximum likelihood estimates.

Finding MLEs by Differentiation

You'll learn: How to use calculus to find the parameter values that maximize the likelihood function and produce maximum likelihood estimates.

Core Idea

Maximum Likelihood Estimation finds the parameter value that makes your observed data "most likely" to have occurred. In the previous lesson, you learned what an MLE is—now we learn how to find it using differentiation.

The key insight: to maximize a function, take its derivative, set it equal to zero, and solve. Since the log-likelihood is easier to work with (products become sums), we differentiate that instead. The maximum of the likelihood and log-likelihood occur at the same parameter value.

The Process

  1. Write the log-likelihood function ℓ(θ) = log L(θ)
  2. Differentiate with respect to θ: compute dℓ/dθ
  3. Set the derivative equal to zero: dℓ/dθ = 0
  4. Solve for θ: this gives your MLE, θ̂
  5. Verify it's a maximum (optional): check the second derivative is negative

Example Intuition

Imagine you're on a hill (the log-likelihood surface). You want to find the peak. Differentiation tells you the slope at any point. Where the slope equals zero, you've found a flat spot—likely the top! You climb by adjusting your parameter estimate until the slope vanishes.

For instance, if you have n observations from a normal distribution with unknown mean μ and known variance, the log-likelihood involves μ. Differentiating with respect to μ and setting to zero yields μ̂ = the sample mean—exactly what intuition suggests!

Key Takeaway: The MLE is found by differentiating the log-likelihood function, setting the derivative to zero, and solving for the parameter—calculus transforms the conceptual "most likely value" into a concrete computational recipe.