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
- /
- Neural Network Detailed Documentation

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
# Neural Network Detailed Documentation

Options

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

2 weeks ago

Hey, I have been trying to find a good source on how to use neural network by utilizing SAS Guide, and SAS Miner(I already have these tools) but I have failed to find a good, and indetailed one. I have data that is consist of 15 independent variables with 30 million observations for each one. The hardware is very weak (8gb ram). Should I even try neural network with SAS tools? If yes:

- How should I preprocess the data?
- How should I train, test, validate the data?
- Even more basically, how should I read the sas output?
- How to decide what are the optimal settings, and why the other settings are not as optimal as the selected ones for the current data?

So I have so many questions for you guys to answer. It would be easier to read it off an article. Is there any good, and indetailed document that cover all these from basics to advanced points in SAS? I would be really happy if I can get any help. Thanks a lot in advanced.

Accepted Solutions

Solution

2 weeks ago

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

Posted in reply to atessu

2 weeks ago

The SAS Enterprise Miner Reference Help (try this link: http://go.documentation.sas.com/?docsetId=emref&docsetTarget=p0zbgj1tu3h1uhn1x6regixbdg7v.htm&docset...) has a lot of information including a section on **Preparing the Data** and many other details about how neural networks work.

If the link above doesn't work, start here:

http://support.sas.com/documentation/onlinedoc/miner/index.html#minercurrent

Click on

Then look under Node Reference: Model Nodes > Neural Network Node : Reference

All Replies

Solution

2 weeks ago

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

Posted in reply to atessu

2 weeks ago

The SAS Enterprise Miner Reference Help (try this link: http://go.documentation.sas.com/?docsetId=emref&docsetTarget=p0zbgj1tu3h1uhn1x6regixbdg7v.htm&docset...) has a lot of information including a section on **Preparing the Data** and many other details about how neural networks work.

If the link above doesn't work, start here:

http://support.sas.com/documentation/onlinedoc/miner/index.html#minercurrent

Click on

Then look under Node Reference: Model Nodes > Neural Network Node : Reference

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

Posted in reply to atessu

2 weeks ago

Thanks a lot. This document is what I am looking for. Do you think I should give it a try for neural with (8gb) ram 30M observations for each 30 independent variables? I can small down the data but I am not sure if that would have been good enough for training.

I am trying to estimate the probability of default of a customer with many independent variables such as credit card limits, demographic informations, and etc.

I am trying to estimate the probability of default of a customer with many independent variables such as credit card limits, demographic informations, and etc.

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

Posted in reply to atessu

2 weeks ago

Yes, you can definitely give it a try - you might want to do Variable Selection first, then you can try either the Neural Network node or HP Neural from the HPDM tab to take advantage of parallel processing.