Cannot connect ReadCSV to Learner

quocblequocble Member Posts: 5 Contributor II
edited November 2018 in Help
Hi,

Using Rapid-i 5.0  / Win7 -

I'm trying to bypass the whole import process and just read the CSV directly, label it, then pass to the learner (Bayes).  When I connect ReadCSV directly to the results, i see the result as expected with the label matched (to column 1 in my case).    But when I connect the readcsv to the Bayes, it keeps saying that my exampleset do not have a Label.  I've gone through the forum and found that most people are using CSVExampleSet which I can't find anywhere on the Operators. Is that deprecated as part of 5.0?    And When I edit the XML directly, and I click the checkmark, the program doesn't seem to read the new xml source code.   

Thank you in advance.  Btw. Rapid-i is a really amazing platform , comphrensive & fast.!!

-Quoc

Answers

  • CleoCleo Member Posts: 44 Maven
    Hello quocble,

    I ran into a similar problem when I first started with RM. 
    http://rapid-i.com/rapidforum/index.php/topic,1285.0.html

    Also FYI RM offers some useful courses:
    http://rapid-i.com/component/option,com_virtuemart/page,shop.browse/category_id,8/Itemid,180/

    Hope this helps,
    Cleo
  • quocblequocble Member Posts: 5 Contributor II
    Did you import the data every time?  I've successfully tested with that method already. I just want to by pass the import process altogether....
  • CleoCleo Member Posts: 44 Maven
    Yes I import the data that I want RM to use every time.  As far as I know all learners would require some input data to make predictions with.
  • quocblequocble Member Posts: 5 Contributor II
    It is pretty obvious that supervised learning requires training data unless it's an unsupervised type of learning. 

    I was going to feed training data, but directly from CSV on the drive, instead of the Rapid Miner's Repository.  I'm just trying to automate as much as possible..

    -Quoc
  • landland RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 2,531 Unicorn
    Hi,
    thank you Quoc for your kind words.

    First of all the CSVExampleSource operator is still part of RapidMiner, but take a look at it's name: It urgently needed to be renamed :) It now simply is names Read CSV, which exactly says what it does. And it seems, that you have already found it :)

    The second problem with the XML editing: I always make the same mistake, since I'm used to RapidMiner 4, but there's now a small button in the icon bar on top of the xml. You now have to press the green "apply" button, to reflect changes in XML inside the process. But this should work, at least it does in my version.

    To the repository: It's strongly recommended to use the repository, at least during design time of the process, since it provides the meta data of your data files, so that all the nice checkings, preconditions and quickfixes can be applied. If you are in need of automation, you might do the import into the repository inside a process...

    So much from me,
      Sebastian
Sign In or Register to comment.