Part of Advances in Neural Information Processing Systems 8 (NIPS 1995)
Bambang Parmanto, Paul Munro, Howard Doyle
Central to the performance improvement of a committee relative to individual networks is the error correlation between networks in the committee. We investigated methods of achieving error indepen(cid:173) dence between the networks by training the networks with different resampling sets from the original training set. The methods were tested on the sinwave artificial task and the real-world problems of hepatoma (liver cancer) and breast cancer diagnoses.