|
Title:
|
ELIMINATING BORDER INSTANCES TO AVOID OVERFITTING |
|
Author(s):
|
Khalil El Hindi , Mousa Al-akhras |
|
ISBN:
|
978-972-8924-87-4 |
|
Editors:
|
António Palma dos Reis |
|
Year:
|
2009 |
|
Edition:
|
Single |
|
Keywords:
|
Artificial Neural Network Training, Machine Learning, Overfitting, Over Learning, Noise Filtering, Instance Reduction. |
|
Type:
|
Full Paper |
|
First Page:
|
93 |
|
Last Page:
|
99 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
This work addresses the problem of overfitting the training data. We suggest smoothing the decision boundaries by
eliminating border instances from the training set before training neural networks. Our empirical results show not only a
reduction in the number of training epochs but also a significant improvement in classification accuracy. |
|
|
|
|
|
|
