Oops! I flew away too fast. :-)
[Just repeating my reply in MSN DM group to share]
If you're going to understand the student profiles in general, try to build
a clustering model to see what distinct groups are there. If you want to
understand relationship between the attributes (columns), you could build
decision trees to predict everything from everything. Browing the tree
content and dependency net will give you understanding on how they are
related each other. Typically, data mining process also starts with certain
business problems to solve, say, optimizing a business practice. For
instance, how can we minimize books lost. You can prepare all relevant data
for books being lost and create a decision tree model. From the rules in the
tree, you will be able to see what attributes contribute
positively/negatively to losing books. This is just an example from my head.
I'm sure there will be tons of such in real world. Note that you could use
DTS to prepare your data ready for data mining.
--
Peter Kim
This posting is provided "AS IS" with no warranties, and confers no rights.
[quoted text, click to view] "Peter Kim [MS]" <peterkim@online.microsoft.com> wrote in message
news:OmL$Kf4CEHA.240@tk2msftngp13.phx.gbl...
> [Just repeating my reply in MSN DM group to share]
>
> --
> Peter Kim
> This posting is provided "AS IS" with no warranties, and confers no
rights.
>
> "smen" <anonymous@discussions.microsoft.com> wrote in message
> news:F755B373-1FF9-4AED-8638-A6B4D148AD38@microsoft.com...
> > Dear all,
> > I'm a student whose doing extensive research on data warehousing and
data
> mining. With me, is a full dataset of students' record from a local
> university. Information's which is in the dataset consists of students
> subjects, grades, personal information, library information, to hostel
> information to finance information of each particular student. Basically,
I
> have the entire dataset at my disposal. Lucky me. Since I'm a rookie in
the
> field of data warehousing and data mining, can the good people of this
forum
> share your ideas with me on what kind of interesting information that I
can
> extract from this sort of dataset. What type of interesting patterns
should
> I look for in a student's database. From the feedback given by you, I'll
> create questionnaires to further expend.
> >
> > Thank in advance for the help people.
> >
> > p/s: I'll post in this forum regularly to update those who are working
or
> in the same interest.
> >
> > smen
>
>
