topic Can JMP do causal forecasting? in Discussions

Can JMP do causal forecasting?

priyanka — Tue, 24 Apr 2012 12:26:45 GMT

Hi all,

I have been able to generate the baseline numbers with the tool using variants of forecasting models - ARIMA and smoothing models and automate that part as well. But I am very keen on knowing if JMP/ JSL allow causal forecasting. If yes, then how?

Please help.

Regards,

Re: Can JMP do causal forecasting?

priyanka — Sun, 29 Apr 2012 07:57:48 GMT

Can someone please help

Re: Can JMP do causal forecasting?

mfisher — Sun, 29 Apr 2012 13:24:24 GMT

You can do causal forecasting in JMP by employing a transfer function within the Time Series Analysis platform. Within Time Series Analysis, your independent (causal) variable goes into the Input List box, the time series variable is in Y, and the date or time in Time ID. The next window is the Transfer Function analysis window, with diagnostics. Click the red "hot spot" triangle and select Transfer Function. Choose the appropriate error terms for your time series variable on the left. On the right, the independent variable can be lagged, or set to have an effect over several time periods, but for a simple concurrent model leave those parameters at zero. Transfer functions can become quite complicated. Click Estimate, and the next window provides output and diagnostics.