Title :
Boosting Learning Machines with Function Compositions to Avoid Local Minima in Regression Problems
Author :
Zegers, Pablo ; Correa, Gonzalo
Author_Institution :
Univ. of the Andes, Santiago
Abstract :
We present an improved cascaded learning method that is based on mathematical functions compositions, instead of additive models as normally done in boosting approaches. Regression experiments are done to support the usefulness of this architecture and training procedure. The method allows to produce a strong learner with increased probability of avoiding local minima.
Keywords :
learning (artificial intelligence); regression analysis; support vector machines; SVM; boosting learning machine; cascaded learning method; mathematical functions composition; probability; regression problem; Boosting; Educational institutions; Guidelines; Learning systems; Least squares methods; Machine learning; Mathematical model; Neural networks; Size measurement; Support vector machines;
Conference_Titel :
Neural Networks, 2007. IJCNN 2007. International Joint Conference on
Conference_Location :
Orlando, FL
Print_ISBN :
978-1-4244-1379-9
Electronic_ISBN :
1098-7576
DOI :
10.1109/IJCNN.2007.4371263
