Presentation | 2002/10/11 Size Reduction of Neural Network Trees through Retraining Takaharu TAKEDA, Qiangfu ZHAO, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | There are mainly two approaches for machine learning. One is the symbolic approach and another is the non-symbolic approach. Decision tree (DT) is a typical model for symbolic learning, and neural network (NN) is the most popular model for non-symbolic learning. Neural network tree (NNTree) is a DT with each non-terminal node being an expert NN. NNTree is a learning model that may combine the advantages of both DT and NN. Through experiments we found that the size (number of nodes) of an NNTree is approximately proportional to the number of training data. Thus, we can reduce the tree size by using less training data. This, however, will also reduce the performance of the system. In this paper, we propose to reduce the size through training with partial data, and then compensate the reduction in performance through retraining. Using this method, it is possible to reduce the tree size and keep the the performance unchanged. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Neural network / decision tree / genetic algorithm / neural network tree |
Paper # | NC2002-58 |
Date of Issue |
Conference Information | |
Committee | NC |
---|---|
Conference Date | 2002/10/11(1days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | |
Vice Chair | |
Secretary | |
Assistant |
Paper Information | |
Registration To | Neurocomputing (NC) |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Size Reduction of Neural Network Trees through Retraining |
Sub Title (in English) | |
Keyword(1) | Neural network |
Keyword(2) | decision tree |
Keyword(3) | genetic algorithm |
Keyword(4) | neural network tree |
1st Author's Name | Takaharu TAKEDA |
1st Author's Affiliation | The University of Aizu Graduate School of Computer Science and Engineering() |
2nd Author's Name | Qiangfu ZHAO |
2nd Author's Affiliation | The University of Aizu Graduate School of Computer Science and Engineering |
Date | 2002/10/11 |
Paper # | NC2002-58 |
Volume (vol) | vol.102 |
Number (no) | 382 |
Page | pp.pp.- |
#Pages | 6 |
Date of Issue |