Data mining from multiple heterogeneous relational databases using decision tree classification

Nowadays, the expansion of computer networks and the diversity of data sources require new data mining approaches in multi-database systems. We propose a classification approach across multiple heterogeneous relational databases. More specifically, given a set of inter-related databases, we use a regression model for predicting the most useful links that will be connected to build a multi-relational decision tree. Experiments performed on different real and synthetic databases were very satisfactory compared with previous classification approaches in multiple databases.

► We present a classification approach across multiple heterogeneous relational databases.
► We use a regression based SVM model for predicting the most useful links.
► Then, we perform a multi-relational decision tree classification algorithm.
► Experiments on real and synthetic databases were very satisfactory.