= Likelihood ratio
{title2=$\Lambda(x)=f_1(x)/f_0(x)$}
For two specified model densities or mass <functions> $f_1,f_0$, the <likelihood ratio> at observed data $x$ is $\Lambda(x)=f_1(x)/f_0(x)$ wherever the denominator is positive. It <measures> relative support for these two models at that observation, not a posterior <probability> without prior odds. One Bernoulli success gives ratio $0.8/0.2=4$ for success <probabilities> $0.8$ and $0.2$; a failure gives ratio $0.2/0.8=1/4$. For independent observations the ratios multiply. In a <likelihood-ratio test> of composite models, optimized <likelihoods> replace the two fixed densities, with an explicitly stated numerator convention.
Back to article page