turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Find a Community

- Home
- /
- Analytics
- /
- Data Mining
- /
- How to see correlation matrix in sas enterprise mi...

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

Highlighted
# How to see correlation matrix in sas enterprise miner

Options

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

05-14-2017 10:45 AM

Hi,

I wonder if you can please help me?

I am trying to perform some algorithms (Neural Network, SVM, Decision Tree and Random Forest) on my dataset.

What I am trying to see is how can I get Correlation Matrix, Confusion Matrix when I apply Model Comparison Node.

Basic Flow of my diagram is as follow

Import >>>> Stat Explorer >>>> Data Partition >>>> ANN, SVM, HP Forest, DT>>>> Control Point >>>> Model Comparison

Could you also please suggest any suitable node or critaria which may give me more results with these 4 algorithms in terms of best model comparison and results.

Regards

Accepted Solutions

Solution

05-20-2017
02:49 AM

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

Posted in reply to geniusgenie

05-18-2017 02:40 PM

Hello,

Great Question.

A correlation matrix can be obtained using the variable clustering node.

Try attaching the model comparison node to the variable clustering node and then run the variable clustering node. From the results screen, click view > Model > Variable Correlation.

To surface the numeric values used to derive the graph, click on the Table shortcut button at the top right of the results screen.

An Event Classification Table is provided in the output window of the results in the Model Comparison node, which provides you the TN, FN, TP and FP counts.

With respect to the additional graphs or results, I find the model comparison node does a pretty good job. I really enjoy examining the Lift chart provided by the model comparison node because the lift charts gives us a graphical representation of each model's performance over the prior, that is a random guess.

I hope this helps!

Best,

Robert

All Replies

Solution

05-20-2017
02:49 AM

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

Posted in reply to geniusgenie

05-18-2017 02:40 PM

Hello,

Great Question.

A correlation matrix can be obtained using the variable clustering node.

Try attaching the model comparison node to the variable clustering node and then run the variable clustering node. From the results screen, click view > Model > Variable Correlation.

To surface the numeric values used to derive the graph, click on the Table shortcut button at the top right of the results screen.

An Event Classification Table is provided in the output window of the results in the Model Comparison node, which provides you the TN, FN, TP and FP counts.

With respect to the additional graphs or results, I find the model comparison node does a pretty good job. I really enjoy examining the Lift chart provided by the model comparison node because the lift charts gives us a graphical representation of each model's performance over the prior, that is a random guess.

I hope this helps!

Best,

Robert

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

Posted in reply to RobertBlanchard

05-20-2017 02:51 AM

Thanks a lot Robert and sorry for late reply.

I have followed your suggestion and it works perfectly for me..

Thanks again

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

Posted in reply to RobertBlanchard

Tuesday

BIG THANKS!