I have developed an ANN for some experimental results. However, because of the test procedure, the results are destroyed within the first 30 seconds. Furthermore, the results are recorded in 24 hours of employing the test.
I want to know how I can use the trained network to determine the results at the first stage and to predict the results after 24-hour.
Hi Mahboobeh,
Thank you for your response.
I read the article through and I think you mean using the "sim" function to predict the outcomes using the trained network. But, again, when I gave this function an input out of the first input range it gave me some unreasonable results.
Hello Ebrahim,
You are probably making some mistakes in normalization of the input/output and bringing them back to the original range. If you have trained a network with normalized data, you need to feed the new data in the normalized form, "sim" the output and then bring them into the original range.
By the way, the question is not clear. Take a look at recurrent NNs.
- Badges
- Science topic