Likelihood

A likelihood $\mathcal{L}(x|\theta )$ computed for a data point $x$ with $\theta$ a set of parameter that is fixed, measures how well the statistical model with parameters $\theta$ explains the observed data $x$.

In other words, it tells us what’s the likelihood that the point $x$ belongs to the distribution explained by the model with parameters $\theta$, i.e. it quantifies the likelihood of observing the data $x$ assuming the model’s parameters are $\theta$.

The likelihood is NOT a probability!

---

tags:#probability-theory