EViews.com

Hello all! Season's greetings! I am writing this to know if it is somehow possible in EViews to calculate statistical distribution functions without using commands. I wish to compute CDF, PDF and Quantals for some series of observations for uniform distribution and t-distribution. Being visually challenged, I depend on screen readers, such as Jaws; and unfortunately, EViews has, so far, not paid any attention to improve its compatibility/accessibility with screen readers. When I try to insert command for calculating statistical distribution functions (mentioned above), I get error message (like @cunif is not defined). Please assist me so that I can compute the statistical distribution functions of my choice independently without much trouble!
Regards!
Vikas Dixit

Hello.

I'm not sure I understand what you are asking. Could you rephrase?

Hi! Though I hardly left any scope for ambiguity, let me again repeate my query. On the EViews Help Webpage as well as in the Command Reference Guide, there is a topic under Operators and Functions "Statistical Distribution Functions". It describes how one can calculate the cumulative distribution function (CDF), probability distribution function (PDF), quantals and random numbers. This can be done for various distributions such as beta, uniform, etc with commands like @cunif(x,a,b) for CDF, etc.
My simple query is: can this type of calculation be done without using commands from the main menu (such as we make various other computations/estimations from the main interactive menu)?
I hope my concern is understood this time; I do not want to use commands for calculating statistical distribution functions, rather I wish to do this from the main menu options. Is it possible anyway?
Thanks and regards!
Vikas Dixit

Ah, that makes sense now, thank you.

Unfortunately functions are not available from the menu system - they are, by definition, part of the command language only.

However it would be relatively simple to create an add-in that presented some simple options. I've created on (attached) that does beta, uniform, normal and t-dist. If you want to expand it to other distributions, it should be relatively easy to see how it works.

OK, thank you very much for this idea. However, I am not at all familiar and comfortable in create add-ins; this may annoy me further. Anyway, I will try to practice with one provided by you on Monday on my departmental PC where I have installed EViews. Then I will revert back if I face any problem.
Thanks and regards!
Vikas Dixit

Hi Gareth! I have installed the add-in provided by you and it is now there in the options. However, I wonder if there is any trick to choose the parameters. For example, if I have a series of say 57 observations and I wish to calculate, say the density function for each value of this series under uniform distribution, then your add-in asks for three parameters. My two queries are as under: Do I need to calculate the density for each observation of my series separately by placing the single observation as the first parameter, while minimum and maximum of the series as other two parameters every time? Is it not possible to put the name of the series in place of the first parameter and EViews automatically generates other two parameters (this also applies in case if I choose a different distribution)?
Please take trouble to clarify!
Regards!
Vikas Dixit

The EViews functions you mentioned all require you to specify the parameters of the function.

Let me add something to what Gareth has said.

The parameters you have to supply describe the distribution. The series you have are (presumably) draws from that distribution. So even in the case of the uniform it doesn't follow that the min and max of the distribution are the min and max of the series.