Data Mining Introduction Part 2

Question

Post reply

Data Mining Introduction Part 2

Daniel Calbimonte

SSCertifiable

Points: 6200
More actions
December 31, 2012 at 12:03 am

#268794

Comments posted to this topic are about the item Data Mining Introduction Part 2

Viewing 15 posts - 1 through 15 (of 15 total)

You must be logged in to reply to this topic. Login to reply

davoscollective SSCertifiable Points: 6405 More actions · Answer 1

Great article, I think it will really help people new to data mining models in SSAS.

An important point: you said in the beginning "we need a lot of data to predict the future"

and then in the end of the article reaffirmed that you had predicted the future.

No amount of data will predict the future, and those people developing data mining models have a responsibility to ensure that their clients don't believe that they can magically predict the future. All you can do is make predictions within probability, and recognise that the probability value is only as good as the training data. The more data you feed the model, the closer it will converge on its true probability value which is not the same as converging towards 100% accuracy.

gsamborn Old Hand Points: 314 More actions · Answer 2

Everything was easily understandable until you got to the example using the ProspectiveBuyers table.

You stated: "For this example, we are going to use the table dbo.Prospectivebuyers table that is included in the AdventureWorksDW database. Let’s move to the Mining Model Prediction Tab."

When I got to the "Mining Model Prediction" tab, the only model I see is from the previous example.

Maybe its just me, but you seemed to jump ahead a few steps.

Any clarification would be greatly appreciated since I am NEW to this area.

Thanks.

imSQrLy Ten Centuries Points: 1335 More actions · Answer 3

It is not just you. I went back and read the first article so i would be ready for this second one. I am the same place you are, I only have what we created in the first article.

Any updates to this article that help us get through it would be great!

Jimmy

"I'm still learning the things i thought i knew!"

Daniel Calbimonte SSCertifiable Points: 6200 More actions · Answer 4

I agree.

I used the adventureworks project for SQL 2012 for the part 2.

http://msftdbprodsamples.codeplex.com/releases/view/55330

gsamborn Old Hand Points: 314 More actions · Answer 5

Hi

I forgot to mention the version of AdventureWorks I used was 2008R2.

Daniel Calbimonte SSCertifiable Points: 6200 More actions · Answer 6

The steps should be similar, but some names may change.

gsamborn Old Hand Points: 314 More actions · Answer 7

If you re-read my original post (and the post following that), you'll realize that I'm asking "What steps?".

You seem to jump over a number of steps.

Any help would be greatly appreciated.

angad.singh88 Old Hand Points: 300 More actions · Answer 8

I am getting "Age" as the most important factor to buy or not a Bike. Is there anyone else getting the same? or what am I missing. ?

--Angad Singh
If I Stop Learning, I Cease to Be A SIKH !

angad.singh88 Old Hand Points: 300 More actions · Answer 9

Hello Daniel,

When can I expect the part 3 for Data Mining. ?

Thanks

Angad

--Angad Singh
If I Stop Learning, I Cease to Be A SIKH !

Daniel Calbimonte SSCertifiable Points: 6200 More actions · Answer 10

Daniel Calbimonte

SSCertifiable

Points: 6200

January 23, 2013 at 5:56 am

#1580034

Next month please

Thomas LeBlanc SSCrazy Eights Points: 8085 More actions · Answer 11

Great series. Thanks!!!

Thomas

Thomas LeBlanc, MVP Data Platform Consultant

e.parfenov SSC Rookie Points: 35 More actions · Answer 12

I have followed your instructions on my own installation and in the last step i get the same probability for every potential customer in the table ProspectiveBuyer. Do you have an idea why?

gsamborn Old Hand Points: 314 More actions · Answer 13

gsamborn

Old Hand

Points: 314

March 7, 2015 at 10:45 am

#1782358

Absolutely useless!

Samuel Vella SSCrazy Eights Points: 8055 More actions · Answer 14

Couple of points about this article

1. The Adventure Works Multidimensional Model Project needs to be downloaded (http://msftdbprodsamples.codeplex.com/releases/view/55330). Would be nice if the article was updated to reflect this.

2. The model created in the previous article might look similar but has a HUGE number of small differences when compared with the model used in this article and is pretty much worthless except as a very basic end to end example.