Many probability distributions have unknown parameters; We estimate these unknowns using sample data. The Likelihood function gives us an idea of how well the data summarizes these parameters.
The “parameters” here aren’t population parameters— they are the parameters for a particular probability distribution function (PDF). In other words, they are the building blocks for a PDF, or what you need for parametrization.
Let’s say you’re interested in creating a probability density function that represents binomial probabilities for getting a heads (or tails) in a single coin toss. You’re going to estimate the likelihood of getting heads from your data, so you run an experiment.
If you get two heads in a row, your likelihood function for the probability of a coin landing heads-up will look like this:
If you toss once more and get tails (making HHT), your function changes to look like this:
Although a likelihood function might look just like a probability density function, it’s fundamentally different. A probability density function is a function of x, your data point, and it will tell you how likely it is that certain data points appear. A likelihood function, on the other hand, takes the data set as a given, and represents the likeliness of different parameters for your distribution.
Defining Likelihood Functions in Terms of Probability Density Functions
Suppose the joint probability density function of your sample X = (X1,…X2) is f(x| θ), where θ is a parameter. X = x is an observed sample point. Then the function of θ defined as
L(θ |x) = f(x |θ)
is your likelihood function.
Here it certainly looks like we’re just taking our probability density function and cleverly relabeling it as a likelihood function. The reality, though, is actually quite different. For your probability density function, you thought of θ as a constant and focused on an ever changing x. In the likelihood function, you let a sample point x be a constant and imagine θ to be varying over the whole range of possible parameter values.
If we compare two points on our probability density function, we’ll be looking at two different values of x and examining which one has more probability of occurring. But for the likelihood function, we compare two different parameter points. For example, if we find that L(θ1 | x) > L(θ2 | x), we know that our observed point x is more likely to have been observed under parameter conditions θ = θ1 rather than θ = θ2.
Properties of Likelihoods
Unlike probability density functions, likelihoods aren’t normalized. The area under their curves does not have to add up to 1.
In fact, we can only define a likelihood function up to a constant of proportionality. What that means that, rather then being one function, likelihood is an equivalence class of functions.
Likelihoods are a key part of Bayesian inference. We also use likelihoods to generate estimators; we almost always want the maximum likelihood estimator.
Robinson, E. (2016). Introduction to Likelihood Statistics. Retrieved December 23, 2017 from:
Wasserman, L. (n.d.). Lecture Notes 6 1 The Likelihood Function – CMU Statistics
Retrieved December 23, 2017 from: http://www.stat.cmu.edu/~larry/=stat705/Lecture6.pdf
Zhang, K. (2011). Principles of Data Reduction. In Special Topics in Statistical Theory.
Retrieved December 23, 2017 from: http://www.math.ntu.edu.tw/~hchen/teaching/StatInference/notes/ch6.pdf
Confused and have questions? Head over to Chegg and use code “CS5OFFBTS18” (exp. 11/30/2018) to get $5 off your first month of Chegg Study, so you can understand any concept by asking a subject expert and getting an in-depth explanation online 24/7.
Comments? Need to post a correction? Please post a comment on our Facebook page.