I have a distribution of non-normal data. How do I calculate the 60% confidence interval about the mean in Matlab?

Exactly what I needed to hear. Thanks for clearing that up for me!

<t>Thanks again. The model is to be used to forecast currency risk for large government expenditures of various accounts, e.g., Capital, Operations, etc. The problem is that for some periods zero values can occur as nothing was spent. They, therefore, must also be forecastable. I was planning on usi...

Roger, thanks. I also wasn't sure as to whether replacing zeros with next year's values was a valid approach.

<t>It's a time series of expenditures. As an example ...0,0,0,55,4545,1233,634,0,36,0,1234,8636,...etc.Seven years of data with a period of 52 weeks. I was told, for the first three zero values, to replace them with next year's values in the same period. Is that correct? What about autocorrelation i...

<t>Prior to developing a model, my series has zeroes at the start (2 values) and interspersed throughout. The series also expresses seasonality with period 52 weeks. Someone told me to fill in the first two values with the values from the same period in the following year. Not sure if that's valid. ...

Does anyone have a good reference for the Theil U Test that is available as a download? I have seen a number of sources and they each show a dfifferent calculation.Thanks.

In modelling the return series, how far back is reasonable in order to retain the leptokurtic behaviour of returns? Does it approach Normality after a certain amount of time?

Thanks Diana. I'll look it up and don't worry, I'm fairly new at this too.

<t>Well for one thing it's embedded in a Monte Carlo sim, which means re-calculating the MLE each day for each iteration, and there are 10,000 iterations. I remember reading an article about re-calculating them every 10 trading days. Maybe I'll go that route if it's programmable through VBA in Excel...

<t>Thank you for replying. I'm trying to forecast foreign exchange rates out to 12 months on a daily basis. I have built a GARCH(1,1) model from 6 years of daily returns with the GARCH coefficients specified through maximizing the log likelihood of the Student's T-Distribution. The forecast is deter...

Does anyone have any insights into daily forecasting of exchange rates using time-varying GARCH parameters or just modelling time-varying GARCH parameters. I know it can easily be done using constant parameters, but I find that the predictive power loses strength rapidly.

For example, the CAD/USD rate is decreasing. How would you model any change in, say 1 year?

Can someone give me an example?

<r>Does anyone have a simple example (preferably in Excel) where an ARMA model has been fitted to a log transformed and differenced series and then the original, untransformed series has been forecasted?I have been working on a Log, Diff of 12, Diff of 1 and mean subtracted series and now want to fo...