narxnet early prediction problem

1 次查看（过去 30 天）
Victor Plaza
Victor Plaza 2013-4-25
评论： Morteza Hajitabar Firuzjaei 2018-1-25
I have developed a NarxNet to predict the next High on a stock market asset. When i use the early prediction the network only can predict the y(t) once i have provided two delays with its targets and a new value with a target too. The problem is that at the time of using it in real time i can't have the last target, i mean the value to be predicted target so i get the last one y(t-1) and a NaN for the next.
How can i use then the early prediction to get a prediction of the next value?
Thank you for your help
  1 个评论
Morteza Hajitabar Firuzjaei
Morteza Hajitabar Firuzjaei 2018-1-25
Dear Victor,
You can use close loop to predict next values in narxnet, for example:
% one-step-ahead prediction
perf = perform(net,Ts,Y);
[Xs1,Xio,Aio] = preparets(net,inputSeries(1:end-delay),{},targetSeries(1:end-delay));
[Y1,Xfo,Afo] = net(Xs1,Xio,Aio);
[netc,Xic,Aic] = closeloop(net,Xfo,Afo);
[yPred,Xfc,Afc] = netc(inputSeriesVal,Xic,Aic);
multiStepPerformance = perform(net,yPred,targetSeriesVal);
view(netc)
also this information aren't efficient
Beat Regards,
Morteza Hajitabar Firuzjaei

请先登录，再进行评论。

请先登录，再回答此问题。

回答（0 个）

类别

在 Help Center 和 File Exchange 中查找有关 Sequence and Numeric Feature Data Workflows 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