neural nets: validation vs testing?

2016 3 24
2 个回答
回答已采纳
更新时间：2016 3 25
15 次查看（30 天）

0 个投票

wondering what the real difference is between validation and testing? both are tests of the net, it would seem
thanks

请先登录，再回答此问题。

 采纳的回答

Greg Heath
Greg Heath 2016-3-25

2 个投票

NOTE THE DIFFERENCE BETWEEN DESIGN AND TRAINING
1. Data Division
DATA = TRAINING + VALIDATION + TESTING + UNSEEN
2. Design
DATA = DESIGN + NONDESIGN
DESIGN = TRAINING + VALIDATION
NONDESIGN = TESTING + UNSEEN
3. Training
DATA = TRAINING + NONTRAINING
NONTRAINING = VALIDATION + TESTING + UNSEEN
4. Training data is used to directly modify weight and bias values. Performance estimates obtained from training data are HIGHLY BIASED because the same data is directly used for both modifications and evaluation.
5. Validation data is used for
a. ValStopping: Stops training when nontraining
validation error increases for m consecutive
epochs. This enhances the network's ability to
generalize to nondesign ( i.e., testing and
unseen) data. The MATLAB default is m = 6.
b. Model ranking: Ranks multiple designs w.r.t.
performance. Performance estimates are SLIGHTLY
BIASED because validation data indirectly
affects design.
6. Testing data is used to obtain UNBIASED ESTIMATES OF NONTRAINING (including UNSEEN) DATA.
Hope this helps.
Thank you for formally accepting my answer
Greg

更多回答（1 个）

Walter Roberson
Walter Roberson 2016-3-25

0 个投票

The test set may be used as feedback to refine the neural network. The validation set may not be.

类别

在 帮助中心 和 File Exchange 中查找有关 Deep Learning Toolbox 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by