seas dummy bimonthly s=24, fortnight s=26...

[ch2324]

i would like this
(@expand(@mod(@trend,24)+1)=9 take 1 for 01/05/1973 and 01/05/1974.....etc
(@expand(@mod(@trend,24)+1)=10 take 1 for 16/05/1973 and 01/05/1974...etc
....and so on

[EViews Gareth]

That's like saying you would like:
@sqrt(9)=4

What you're describing is mathematically impossible. You can't change the definition of mathematical functions.

Again, I think you need to ask yourself whether the order of the dummy variables really matters or not.

[ch2324]

yes, but for the monthly data it make regular dummy:
1973m5 = @month=5 take 1
1973m6= @month=6 take 1
...etc
why for bimonthly not? s=26 it should start from 5th month ( the 1st bimonthly of May)

[ch2324]

i hope that you understand me

[EViews Gareth]

I understand what you're saying. I'm telling you you are wrong.

[ch2324]

i understand the logic of code, many thanks Mr Gareth. :D

[ch2324]

How can extract all the dummy as series from the show @expand(@mod(@trend,24)+1).

[EViews Gareth]

Don't understand the question.

[ch2324]

extract all 24 series dummy from the group g1

[EViews Gareth]

Extract in to what?

[ch2324]

to use some of them in regression, every time you will use manually @drop(1) to @drop(10) and some others, can i dropping a sequence?
y @expand(@mod(@trend,24)+1, drop(1)............drop(10))
you understood me.

[EViews Gareth]

Just use the group. Or if you only want some of them use g1(1) to refer to the first entry in the group etc....

Or if you are only using a couple of them, don't use the @expand at all, but type out the individual expressions in the regression.

[ch2324]

Great! :D thanks! Mr Gareth.

[ch2324]

(*) EV 9 Users Guide II, page 29.

Code: Select all

@expand(sex, region, @drop(0,"West"), @drop(1,"North")

add parenthesis at the end of the code ")".
(**) EV 9 Object Ref, adjust table of contents.
(***) EV 9 Help, Statistical Distribution Functions, adjust the right table for some distribution functions.
:)

[EViews Rebecca]

What specifically do you see wrong with (***)?