Likelihood function

A function of the model parameters, with observed data fixed, giving the probability of the data under those parameters: L(θ|x) = P(x|θ). The maximum likelihood estimator is the parameter value that maximizes this function. It is central to both frequentist and Bayesian inference.