Not sure what processes to use or how to approach

ngreisman

Hi,

New to RapidMiner [and data mining] and forum and not even sure what to search for and would appreciate any assistance.  I have watched the videos and read all of the material I can find.

I have multiple files to analyze. 
Each file contains a header row and the same format of data in each subsequent row
The different files contain the output data from running the real-time process that includes both raw data and calculated data.
The data [stored in the rows] was captured during a real-time process with the desired outcome in certain columns of the last row.
I want to use RapidMiner to:
  help me understand the data better
  predict outcomes
  eventually to be used in real-time for decision making

Can I use RapidMiner to help me analyze my data?  If yes, how?

I would really appreciate any help you can give.

Thanks,

Neil

Flake

Very general comments:

- RM does do a lot of things, and probably most of your requirements can be met.
- You may start with tutorials (http://rapid-i.com/content/view/189/212/lang,en/) to know better in order to actually elaborate your questions in more detailed manner for others to help you.

ngreisman

Hi,

Thanks for your suggestion but I watched the tutorials before I posted my question.

In the neural network world, I believe I would send each one of the files though a NN to train it.  Each of the files containing multiple rows of data and an outcome [I think you call it a label].  I don't know how to do this in rapidminer.

Is this a better question?

Thanks,

Neil

Flake

Hi Neil,

Not familiar with NN.

But maybe it may help for you to think about what kind of process you want to perform. Classification or clustering or regression,etc?

Once you first figure out a high-level structure of your idea, you can start with some basic tests already.

Have you checked out the video on these two sites: http://www.neuralmarkettrends.com/index.php/tutorials and http://vancouverdata.blogspot.com/
These are very helpful once you get a general idea about RapidMiner by reading the official manual and watching the official videos.

Hope this may help a bit.

Cheers.

ngreisman

Thanks,

I'll take a look.

Neil

ngreisman

Ok,

I watched all of the videos and it still doesn't answer my question but maybe I can be more specific.

All of the examples I have seen only use 1 input file that contains rows of data. Each row being 'complete'. My data consists of multiple input files, each containing information gathered as a process progressed. The final row contains the information that  I need predicted.

Think of collecting all of the moves of a stock during the day and predicting what the next value would be. Now, instead of analysing one daysnworth of data, I have hundreds (or thousands) to use for training.

As before, any assistance would be appreciated.

Thanks,

Neil