Abstract: Method and server for determining a spam prediction error parameter for a spam prediction parameter are disclosed. The method includes: receiving a plurality of emails destined to a plurality of users where a given email has a spam prediction parameter and a user-interaction parameter indicative of whether an associated recipient of the plurality of users agrees with the spam prediction parameter, and clustering the plurality of emails into at least two clusters having respective subsets of emails. For a given cluster the method includes determining a ground truth parameter by analyzing its subset of emails and the associated user-interaction parameters, and assigning the ground truth parameter to the given cluster. For the given email, the method includes generating the spam prediction error parameter based on a difference between the spam prediction parameter and the ground truth parameter, and storing the spam prediction error parameter in association with the given email.

Abstract: Method and server for determining a spam prediction error parameter for a spam prediction parameter are disclosed. The method includes: receiving a plurality of emails destined to a plurality of users where a given email has a spam prediction parameter and a user-interaction parameter indicative of whether an associated recipient of the plurality of users agrees with the spam prediction parameter, and clustering the plurality of emails into at least two clusters having respective subsets of emails. For a given cluster the method includes determining a ground truth parameter by analyzing its subset of emails and the associated user-interaction parameters, and assigning the ground truth parameter to the given cluster. For the given email, the method includes generating the spam prediction error parameter based on a difference between the spam prediction parameter and the ground truth parameter, and storing the spam prediction error parameter in association with the given email.