Go Premium for a chance to win a PS4. Enter to Win

x
?
Solved

Is historical data always needed for predictions?

Posted on 2013-01-02
2
Medium Priority
?
291 Views
Last Modified: 2016-02-18
Do I always need historical data to make a data mining prediction? Do you have any authoritative reference link on this?

If we don't always need the historical data, then would the following case be realistic for determining the prevailing value or a threshold?

Say Inputs: A, B, C, D, where: C = A/B
Predict: C

Since C is a computation and already exists, what would the Predicted C imply or mean in other words? Do you have an authoritative link on this?
0
Comment
Question by:rodqxo
  • 2
2 Comments
 
LVL 51

Accepted Solution

by:
Mark Wills earned 1500 total points
ID: 38739389
It depends on your data mining model...

Some models are based on historical trends, or need a quorum of data to represent an adequate population sample. For example, regression analysis.

Other data mining models are more disclosure and discovery based on relationships, called association. For example basket market analysis

But for predictions, we do need some kind of regression analysis, so yes, that would normally mean historical data.

Have you seen or walked through the tutorials ? http://technet.microsoft.com/en-us/library/bb677206.aspx 

There is also a reasonable albeit high level explanation of some techniques : http://www.dataminingtechniques.net/data-mining-tutorial/

There is a plethora of information out there - but I think an assumption of volume is not really disclosed as to how much (or more accurately how little is needed).

There are tools (like the graphs) and statistical measures (like confidence interval) that do help qualify the results, and obviously the bigger the sample, the greater the confidence of prediction.
0
 
LVL 51

Assisted Solution

by:Mark Wills
Mark Wills earned 1500 total points
ID: 38739482
Oh, and is SSAS / data mining you can use lift charts to measure : http://msdn.microsoft.com/en-us/library/ms175428.aspx

There is quite a bit in MSDN about Data Mining, trick is learning where the information is hidden (*sigh* typical MS).

For example, a discussion about Associations (your example) then http://msdn.microsoft.com/en-us/library/ms174916.aspx and check out the topics on the left...
0

Featured Post

VIDEO: THE CONCERTO CLOUD FOR HEALTHCARE

Modern healthcare requires a modern cloud. View this brief video to understand how the Concerto Cloud for Healthcare can help your organization.

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

An alternative to the "For XML" way of pivoting and concatenating result sets into strings, and an easy introduction to "common table expressions" (CTEs). Being someone who is always looking for alternatives to "work your data", I came across this …
Microsoft Access has a limit of 255 columns in a single table; SQL Server allows tables with over 255 columns, but reading that data is not necessarily simple.  The final solution for this task involved creating a custom text parser and then reading…
Using examples as well as descriptions, and references to Books Online, show the different Recovery Models available in SQL Server and explain, as well as show how full, differential and transaction log backups are performed
Via a live example, show how to backup a database, simulate a failure backup the tail of the database transaction log and perform the restore.

876 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question