2013 Fiscal Year Final Research Report
Ensemble Learning with Negative Correlation and Complexity Estimation
| Project/Area Number |
23500282
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Research Field |
Sensitivity informatics/Soft computing
|
|---|
| Research Institution | The University of Aizu |
|---|
Principal Investigator |
LIU YONG 会津大学, コンピュータ理工学部, 准教授 (60325967)
|
|---|
| Project Period (FY) |
2011 – 2013
|
| Keywords | ニューラルネットワーク |
|---|
| Research Abstract |
Balanced ensemble learning method has been implemented from negative correlation learning. Balanced ensemble learning is able to create weak learning models by changing the learning error functions with the shifted target values. Such learning functions cannot only help in controlling the complexity of the neural network ensembles, but also improve the accuracy of classifications of the learning models.
In comparison, negative correlation learning is able to generate strong learning models. An novel transition learning method between balanced ensemble learning and negative correlation learning has been proposed. The complexity of the learning models could be controlled during the transition period in the learning process.
|
Research Products
(13 results)
