Is there any way to train a portion of a training data set from a selected dataset for data mini

Hi, all experts here,

I am wondering is there any way to select only a portion of a data set to train the mining model? In this case, I mean we dont need to split the dataset in advance, what I want to do is being able to select any random portion of a selected dataset to train a mining model. Any advices?

I am looking forward to hearing from you and thanks a lot in advance for your advices and help.

With best regards,

Yours sincerely,

With SQL Server 2005 you can replace DSV tables with named queries that are filtered by some condition you specify. One "simple" way to sample, for example is to add a random number column to your source data, and then filter on that random number in the DSV. To replace a table with a named query, simply right click the DSV table and select "Replace with named Query"

We are adding features to the next version of SQL Server Data Mining to assist with scenarios such as these.

|||

Hi, Jamie,

Thanks for your advices. It's been helpful.

With best regards,

Yours sincerely,

Is there any way to blank out certain columns in a single sele

The users doesn't want the data for these 4 columns to print out unless the
formID is one of the ones selected. I can use "" instead. The dataset is
sent to Crystal Report to print out so it's better using "". I think Crysta
l
prints out the word "NULL" if they're set to NULL.
Thanks.
"Raymond D'Anjou" wrote:

> Add a Case for each of these columns.
> Example:
> ...CASE when b.formID in ('2', '16', '11', '12', '1', '13', '10') then NU
LL
> else b.form end as form,...
> "Alpha" <Alpha@.discussions.microsoft.com> wrote in message
> news:9BD3F043-072D-4598-B2BB-BA63529EA142@.microsoft.com...
>
>I don't know anything about Crystal report.
You can use '' for Text datatypes but the numerics and dates may not give
you what you want.
You may have to do a bit of CASTing for these types.
If the users can only choose 1 formID, just build your query differently in
C#.
if formid in ('2', '16', '11', '12', '1', '13', '10') then
select t.tid, t.tdate, t.trip_ticket,t.source,'' AS form, '' AS
formdate, '' AS printed,...
else
select t.tid,t.tdate,t.trip_ticket,t.source,b.form, b.formdate,
b.printed,
"Alpha" <Alpha@.discussions.microsoft.com> wrote in message
news:D27DCE8F-4B22-41C1-BDDF-A344467FCD63@.microsoft.com...
> The users doesn't want the data for these 4 columns to print out unless
> the
> formID is one of the ones selected. I can use "" instead. The dataset is
> sent to Crystal Report to print out so it's better using "". I think
> Crystal
> prints out the word "NULL" if they're set to NULL.
> Thanks.
> "Raymond D'Anjou" wrote:
>|||Hi,Thank you for the reply. I think the system got messed up here. The
"NULL" answer was address to Mesa's question.
My question for you is: Do I just append the "Case..." to the end of my
select
statement?
"Alpha" wrote:
> The users doesn't want the data for these 4 columns to print out unless th
e
> formID is one of the ones selected. I can use "" instead. The dataset is
> sent to Crystal Report to print out so it's better using "". I think Crys
tal
> prints out the word "NULL" if they're set to NULL.
> Thanks.
> "Raymond D'Anjou" wrote:
>|||No, the Case replaces the column name in your Select statement.
So, instead of:
select b.form,...
select CASE when b.formID in ('2', '16', '11', '12', '1', '13', '10') then
NULL else b.form end as form,
"Alpha" <Alpha@.discussions.microsoft.com> wrote in message
news:ED00B94C-E471-4508-9F1C-D8E2803E47EC@.microsoft.com...
> Hi,Thank you for the reply. I think the system got messed up here. The
> "NULL" answer was address to Mesa's question.
> My question for you is: Do I just append the "Case..." to the end of my
> select
> statement?
>
> "Alpha" wrote:
>