The wide applicability of these techniques is also an attractive prospect. The particular
technique studied in this thesis, Negative Correlation Learning [85] can be applied to any
function approximator that can minimise a quadratic error function. Many of the ensemble
methods developed over the last decade of research can and will be applied to new learning
techniques for many years to come. Working at such an abstracted level allows us to exist,
interact and contribute at the junction of several disciplines: artificial intelligence, financial
forecasting, and statistics to name but a few. As the latest machine learning fashions come
and go, so ensemble techniques continue to be applied. From the current en vogue Support
Vector Machines [142, 143], to the more established Hidden Markov Models [52], people
find ways to extract that little bit more information with an ensemble technique.
The wide applicability of these techniques is also an attractive prospect. The particulartechnique studied in this thesis, Negative Correlation Learning [85] can be applied to anyfunction approximator that can minimise a quadratic error function. Many of the ensemblemethods developed over the last decade of research can and will be applied to new learningtechniques for many years to come. Working at such an abstracted level allows us to exist,interact and contribute at the junction of several disciplines: artificial intelligence, financialforecasting, and statistics to name but a few. As the latest machine learning fashions comeand go, so ensemble techniques continue to be applied. From the current en vogue SupportVector Machines [142, 143], to the more established Hidden Markov Models [52], peoplefind ways to extract that little bit more information with an ensemble technique.
การแปล กรุณารอสักครู่..
