Print
Dyma'r cyd-destun: "The relational data mining programmes build hypotheses from huge amounts of rich data in a relational data base format .... This avoids building hypotheses from impoverished data (finite fixed set of descriptors per object) ..."