I finished undergraduate presentation this morning. My speech wasn't smoothly at all. Nervousness made me forget the words, change the pronunciation.
Question: why proposed method perfomed the same as EWMA ? EWMA equation uses only previous value but give good estimation but single-layer neural network see the historical values but cound't give an accurate estimation. I replied that, our neural network model has only one weight table which is adjust due to current inter-arrival time and is used to calculate new estimation. So when the next inter-arrival is longer or shorter than history, it causes high error. It is obvious that single layer model ability is limit in such cases.
Monday, February 06, 2006
Presentation
Posted by Hana at 1:04 PM
0 comments:
Post a Comment