# Data Mining Introduction

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

• Its a nice article Daniel. The only comment I have is that in the Data Mining sub heading, the final paragraph states that you don't need to create a cube or an analysis services project, yet when you get into the tutorial you create an analysis services project so I am a bit confused.

Also, a question I would like to ask would be: "does this example work just as well with the normal adventureworks database as it does with AdventureworksDW?"

Other than that it seems clear enough...

• Hi Daniel,

Very nice article!

At the last examples, shouldn't the probability to buy be 42%? The one you gave is the probability not to buy, right?

Regards,

Luis Torres

• Thank you for an informative introduction into applying data mining models.

---

Note to developers:
CAST(SUBSTRING(CAST(FLOOR(NULLIF(ISNULL(COALESCE(1,NULL),NULL),NULL)) AS CHAR(1)),1,1) AS INT) == 1
So why complicate your code AND MAKE MY JOB HARDER??!:crazy:

My blog: http://uksqldba.blogspot.com
Visit http://www.DerekColley.co.uk to find out more about me.

• Nice introductory article, but I would have liked another case study instead of the plain old BikeBuyer one.

Need an answer? No, you need a question
My blog at https://sqlkover.com.
MCSE Business Intelligence - Microsoft Data Platform MVP

• luis.torres 47140 (11/12/2012)

At the last examples, shouldn't the probability to buy be 42%? The one you gave is the probability not to buy, right?

I came here to point out the same thing but I also have a question about it.

Q. Why does the order change?

A. The answer is that the result is ordered by probability (high to low)

I liked the article too but I think more explanation is needed as to why we are selecting the options indicated. I am quite good at SSAS and SQL in general so I can work it out and expand for myself but I think for new comers it starts well but degenerates slightly in to a "click this, then this, then this" type of tutorial rather than "select this so that you can see this and you do this because...".

e.g: Perhaps explain what a Singlton query is and what the alternatives are. I realise this is explained in BOL but then so is the guts of all of this tutorial. Also why choose PredictHistogram and what are the alternatives?

Apart from this I think its great and look forward to the follow ups.

Dave

From BOL:

Singleton queries do not require an input table; instead, a set of data is passed into the model, from which a single prediction is returned in real time

David Bridge
David Bridge Technology Limited
www.davidbridgetechnology.com

• rob.carter 79958 (11/12/2012) Its a nice article Daniel. The only comment I have is that in the Data Mining sub heading, the final paragraph states that you don't need to create a cube or an analysis services project, yet when you get into the tutorial you create an analysis services project so I am a bit confused.

Data mining projects in SSAS also build a cube as a way of storing the data mining model, but I think the point was that you don't have to create these things first, the data mining wizard does a lot of it for you.

Great article, but I advise caution with people new to data mining. Important items to introduce are the concepts of training, testing and evaluation datasets as well as the danger of over-training your model. I don't believe it's vital to understand the mathematics behind data mining models, it's much more important to have solid business domain knowledge, but I do think that these things should not be treated as a black box.

• does this example work just as well with the normal adventureworks database as it does with AdventureworksDW?

The tables are in the AdventureworksDW database, but they are simple Views.

At the last examples, shouldn't the probability to buy be 42%? The one you gave is the probability not to buy, right?

Correct

Nice introductory article, but I would have liked another case study instead of the plain old BikeBuyer one.

Hehehe, it is hard to have a good sample of data.

I came here to point out the same thing but I also have a question about it.

Q. Why does the order change?

A. The answer is that the result is ordered by probability (high to low)

According to the people characteristics, the probability to buy changes.

• Hi Friend,

Thanks for posting this article it's very useful for my self and you should keep on posting the BI related articles.

Thanks and regards,

Mahalingam M

• Hello, I would like to know when the next article is coming, if not, do you have another article as a follow-up for this one?

Thank you

• Thanks for the replies, I will write about this topic and hopefully you will have a part 2 next month.

• Thanks I will be looking forward to it

• A lot of people are scared off data mining and inferential statistics because of the complex and sometimes counter-intuitive maths. Here is a classic examaple that I'm still struggling to get my head around:

There are 3 sections here.

1. Create a Data Source

2. Create a Data View

3. Create a Data Mining Project

4. Predict information using the Mining Model

...One of the symptoms of an approaching nervous breakdown is the belief that ones work is terribly important.... Bertrand Russell

• Nice article.

• Nice article !!!

Viewing 15 posts - 1 through 15 (of 46 total)