| Discussion: | All Topics |
| Topic: | Neural Networks Equalizer |
| Post a new topic to the General Discussion discussion |
| ||||
| Subject: | Neural Networks Equalizer |
| Author: | ThaiHoa |
| Date: | Nov 6 2003 |
 |
|
Dear Sir,
I’m a forth year University student in Viet Nam. My thesis need
somethings about The Noise Cancellation by using Neural Networks. I know that
you are an expert programmer in Neural Networks using Matlab. I have some
problems in my thesis, one of them is the cancellation noise.
I don’t know why my program is not stable. In some cases, it gives
suitable results, but sometimes it gives wrong results. For example: in some
cases, after training network, we use the new ‘net’ to simulate our data, it
gives some changes suddenly on the waveform (You can see clearly in F1.fig &
F2.fig). In some cases, it can’t study the target waveform.
Here, I use Batch Training, it gives more correct than Incremental
Training. But it takes a lot of time in training & simulation. I have changed
‘sim.m’ to ‘MySim.m’ to reduce memory, but train.m not yet.
Could you show me how to study targets more efficiently? Could you show me why
there are still some changes suddenly & why my program is not stable? If you
agree, may I have some codes related with my problems.
Thanks you very much. I’m looking forward to hearing from you.
Your
faithfully,
DUONG
THAI HOA
+++Duong Thai Hoa +++
| |||||||
| Post a new topic to the General Discussion discussion | |||||||