Trim
Moderators: EViews Gareth, EViews Moderator, EViews Esther
Re: Trim
Also, on an unrelated note, why does my Trim add-in keep disappearing from the add-in menu from time to time? Also, why do my winsorised values differ from the values of the percentiles I have calculated? For example, eviews gives me an upper winsorised value of 414 but the 0.995 percentile is 371 as calculated by excel.
-
- Fe ddaethom, fe welon, fe amcangyfrifon
- Posts: 13315
- Joined: Tue Sep 16, 2008 5:38 pm
Re: Trim
It shouldn't disappear - are you sure you have a series object open?
Hard to say about the numbers without seeing the data.
Hard to say about the numbers without seeing the data.
Follow us on Twitter @IHSEViews
-
- Fe ddaethom, fe welon, fe amcangyfrifon
- Posts: 13315
- Joined: Tue Sep 16, 2008 5:38 pm
Re: Trim
If I want to winsorise and the 99th and 1th percentile at the add-in I have to put in the values 0,1 and 0,99 right?
http://s30.postimg.org/tvan3o2e9/image.jpg
http://s30.postimg.org/tvan3o2e9/image.jpg
-
- Fe ddaethom, fe welon, fe amcangyfrifon
- Posts: 13315
- Joined: Tue Sep 16, 2008 5:38 pm
Re: Trim
Hi,
i installed the add-in Trim. Then i winsorized some variables both in EViews and Stata. Results differ. However, i think Stata result is correct because when i winsorized in Excel, i found Stata result is same as Excel result. So, can you help me?
Thank you
Santosh
i installed the add-in Trim. Then i winsorized some variables both in EViews and Stata. Results differ. However, i think Stata result is correct because when i winsorized in Excel, i found Stata result is same as Excel result. So, can you help me?
Thank you
Santosh
-
- Fe ddaethom, fe welon, fe amcangyfrifon
- Posts: 13315
- Joined: Tue Sep 16, 2008 5:38 pm
Re: Trim
Hi,
I want to winsorize EGPI3. EViews data sheet shows the following code. I think it is only taking care of the upper outliers.
out = @recode(in<=0.00199999999999967,0.00199999999999967,in)
Modified: 1970 2012 // out = @recode(in>=16.748,16.748,out)
Do I need to post data here?
Since I am new user, i dont know if i can upload EView file.
I tried to attache images but i could not attach them.
Thank you.
I want to winsorize EGPI3. EViews data sheet shows the following code. I think it is only taking care of the upper outliers.
out = @recode(in<=0.00199999999999967,0.00199999999999967,in)
Modified: 1970 2012 // out = @recode(in>=16.748,16.748,out)
Do I need to post data here?
Since I am new user, i dont know if i can upload EView file.
I tried to attache images but i could not attach them.
Thank you.
-
- Fe ddaethom, fe welon, fe amcangyfrifon
- Posts: 13315
- Joined: Tue Sep 16, 2008 5:38 pm
Re: Trim
Hi Gareth,
Maybe. But winsorized data shows that it only takes care of only upper outliers.
What i am doing is this: open a variable, Proc, Trimming & Winsorizing, then in the q% lowest (highest) observations i am writing 0.05 (0.95). Is it correct?
If yes, but the Stata results and Excel calculations do not match with EViews winsorization.
It would be great if you can help me.
Thank you.
Maybe. But winsorized data shows that it only takes care of only upper outliers.
What i am doing is this: open a variable, Proc, Trimming & Winsorizing, then in the q% lowest (highest) observations i am writing 0.05 (0.95). Is it correct?
If yes, but the Stata results and Excel calculations do not match with EViews winsorization.
It would be great if you can help me.
Thank you.
-
- Fe ddaethom, fe welon, fe amcangyfrifon
- Posts: 13315
- Joined: Tue Sep 16, 2008 5:38 pm
Re: Trim
Every thing you are doing is correct, and when I do it I get winsored results on both ends. Again, can you provide an example/details?
Follow us on Twitter @IHSEViews
Who is online
Users browsing this forum: No registered users and 21 guests