nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

JMP Features corresponding SAS VARMAX procedure

DaeYun_Kim

Dear All,

One of our customer is using SAS and he wants to convert SAS program to JMP. He is analyzing some marketing data that is related commercial effect to sales. For more details, he is building a model which is called Marketing Mix Model that is evaluating the effects of several campaign commercial cost to increase sales. Marketing Mix Model is measuring the sales increment after executing commercial costs, and the increment of sales varies from 5 weeks to 12 weeks .

He is using SAS VARMAX procedure to analyze these data.

As I searched in our community, there's no feature corresponding SAS VARMAX procedure. But I wish there will be similar features to implement this analysis.

Is there any recommendation or comments?

And one more thing, is there any plan to add this feature upcoming JMP version? It will be helpful to announce to our customer.

Thanks in advance.

DaeYun

Accepted answers

ian_jmp

Probably the closest thing would be to try the 'Transfer Functions' in the Time Series platform. See, for example:

NamesDefaultToHere(1);

dt = Open("$SAMPLE_DATA/Time Series/SeriesJ.jmp");
dt << runScript("Time Series");

This may or may not be sufficient for thier problem. If not, you could always use the SAS integration (which I assume from your question, they might not want to do), or the integration with R and (presumably) Python.

All comments

ian_jmp

Probably the closest thing would be to try the 'Transfer Functions' in the Time Series platform. See, for example:

NamesDefaultToHere(1);

dt = Open("$SAMPLE_DATA/Time Series/SeriesJ.jmp");
dt << runScript("Time Series");

martindemel

This is related to your question: https://community.jmp.com/t5/JMP-SE-Discussions/econometrics-VARMAX/m-p/10879

VARMAX is heavily used in Econometrics and related financial maths. As Finance industry is not JMP's core market I doubt they will invest in this area, espacially as there are SAS procs and also R or Python routines one could access via JSL.

(Side note:Your Marketing Mix Models are used in Commercial Marketing Analysis software. As far I know those models also have two major downsides. They tend to undervalue long-term equity building activities (brand-effects) in favor to the focus on short-term sales. And if you have a media mix you will have bias in favor of short term activities like TV adds compared to long-term adds like monthly magazines.)

jiancao

Just to share what I heard at last week's US Discovery Summit. JMP will be adding a time series platform in JMP15 to analyze and forecast multivariate time series. It is not the same as VARMAX. I haven't seen how it works in JMP as it is still in its early stages of development.

DaeYun_Kim

Thanks Ian.

I have met with our customer and talked each other to find solution of changing from SAS to JMP.

I will find the way with your recommendation.

Thanks again.

DaeYun_Kim

Hi, Martin

Yes, I totallu agree with your thought. Thanks for your idea and sharing your thoughts.

DaeYun_Kim

Thanks Jian.

It would be great news for our customer.

Thanks for sharing.