Systems and methods are described herein for learning an entity's trust model and risk tolerance. An entity's trust score may be calculated based on data from a variety of data sources, and this data may be combined according to a set of weights which reflect an entity's trust model and risk tolerance. For example, an entity may weight data of a certain type more heavily for certain types of transactions and another type of data more heavily for other transactions. By gathering data about the entity, a system may predict the entity's trust model and risk tolerance and adjust the set of weights accordingly for calculating trust scores. Furthermore, by monitoring how entities adjust weights for different transaction types, default weighting profiles may be created that are customized for specific transaction types. As another example, an entity's trust score, as reported to a requesting entity, may be adjusted based on that requesting entity's own trust model, or how “trusting” the requesting entity is.