Integrating global and local boosting
Anastasia-Dimitra Lipitakis, Gerasimos S. Antzoulatos, Sotiris B. Kotsiantis, Michael N. Vrahatis · 2015
Several data analysis problems require investigations of relationships between attributes in related heterogeneous databases, where different prediction models can be more appropriate for different regions. A new technique of integrating global and local boosting is proposed. A comparison with other well known and widely used combining methods on standard benchmark datasets has shown that the proposed technique leads to more accurate results.