How to calculate the accuracy in neural network?

Hi all,

I am new in this field. I have a date set that has 5 features, 680 records and one column as a label either 1 or 0. I took 20 record as a testing set and the other as a training set. Then I made the label column as a target and the other as inputs. I created the net with (patternnet) function. Then I trained the data. After that I test the network with the testing set. When I compare the outputs of the test with the original target of the testing set, it's almost similar. That means when I calculate the accuracy by using (True Positive + True Negative) / The number of the testing data, I will get a high accuracy.

So is that correct? ( I don't know if I could put the code here or not!!)

I am using Matlab. Thanks in advance.

Answer 1

[Greg Heath]

That works. However it is generally more useful to obtain the training, validation and testing error of each class.

If you search for some of my code in both NEWSGROUP and ANSWERS, it may be helpful

 greg patternnet

 Hope this helps.

Thank you for formally accepting my answer

Greg

[Turki]

Thank you for replay.

I'm wondering because the accuracy is so high. it usually comes 96,97,98. That is why I am not sure about the code.

[noo no]

please I need this code? my mail (<mailto:[e-mail]

Answer 2

[noo no]

please I need this code? my mail (<mailto:[e-mail] [e-mail]>)