"Joe" <hortoristic@gmail dot com> wrote in message
news:Ozk%23iNUmGHA.3352@TK2MSFTNGP02.phx.gbl...
>I have a fair bit of Data Warehouse expeierence but am new to Data Mining.
>Our agency tracks sort of medical information on clients. We do these huge
>assesments and ask every type of question there is from "do you ever feel
>like hurting yourself" to "do you think someone has been taking your
>belongings" - as well as all other types of personal information, race,
>age, how much external help is required, ect - and the end result is what
>program this client is best suited for and how much outside govt money will
>be dedicated to this person to cover their needs either in nursing
>facilities or in-home care.
>
> What I'm struggling with is what sort of Data Mining models could I be
> creating to come up with something useful? How do I come up with
> questions that I could then create DM approaches for?
>

"Jamie MacLennan (MS)" <jamiemac@online.microsoft.com> wrote in message
news:44a03b4d$1@news.microsoft.com...
> One thing you can try is simple clustering of responses to group like
> people together. Common groups would likely get common programs, etc.
> Another is outcomes - you can use past history of responses to predict
> what people have recieved to do a "reverse engineering" of the decision
> process, which will provide insight into how results were achieved and
> allow you to implement changes (it's hard to change when you don't know
> where you are, after all). I would use Decision Trees and Naive Bayes for
> that task. This can be an iterative process so you can see how changes
> you implement impact actual behavior.
>
> For long term strategy I recommend looking more into some documentation.
> Berry and Linoff's book "Data Mining for Marketing..." should be of use
> here.
>
> Thanks
>
> --
>
> -Jamie MacLennan
> SQL Server Data Mining
> This posting is provided "AS IS" with no warranties, and confers no
> rights.
> PLEASE POST NEW QUESTIONS AT
> http://forums.microsoft.com/msdn/showforum.aspx?forumid=81&siteid=1
> "Joe" <hortoristic@gmail dot com> wrote in message
> news:Ozk%23iNUmGHA.3352@TK2MSFTNGP02.phx.gbl...
>>I have a fair bit of Data Warehouse expeierence but am new to Data Mining.
>>Our agency tracks sort of medical information on clients. We do these
>>huge assesments and ask every type of question there is from "do you ever
>>feel like hurting yourself" to "do you think someone has been taking your
>>belongings" - as well as all other types of personal information, race,
>>age, how much external help is required, ect - and the end result is what
>>program this client is best suited for and how much outside govt money
>>will be dedicated to this person to cover their needs either in nursing
>>facilities or in-home care.
>>
>> What I'm struggling with is what sort of Data Mining models could I be
>> creating to come up with something useful? How do I come up with
>> questions that I could then create DM approaches for?
>>
>
>