Bayesian Penalty/Boost Function

FavoriteLoadingAdd to favorites

Scoring functions are widely used in our system, e.g. in search engine and recommendation service. In practice, a scoring function may be paired with a penalty/boost function for additional tweaks. Given an entity $E$ to be scored, the final score is of form $f(E)w(E)$, where $f(E)$ is the original scoring function, and $w(E)$ serves as the penalty/boost function. Usually, it is…
Original post: Bayesian Penalty/Boost Function
Source: Kifi