The Altair Community is migrating to a new platform to provide a better experience for you. In preparation for the migration, the Altair Community is on read-only mode from October 28 - November 6, 2024. Technical support via cases will continue to work as is. For any urgent requests from Students/Faculty members, please submit the form linked here

"Using ExcelExampleSource displays

vasuraghavanvasuraghavan Member Posts: 1 Learner III
edited May 2019 in Help
Hello

I am using RapidMiner SVM learning algorithm for learning from examples. The data is in Excel Spread sheet. If I upload the data to a database table (MySQL) and use DatabaseExampleSource, the process works fine. If I change the DatabaseExampleSource to ExcelExampleSource I get an error "Polinomial attributes not supported. Use Binary2Multiclass learner to wrap the learning algorithm?

Why is this so?

How do I use the Binary2MultiClass operator? Should this be wrapping the learning algorithm or the input?

Thanks and regards

Vasu Raghavan
Tagged:

Answers

  • TobiasMalbrechtTobiasMalbrecht Moderator, Employee-RapidMiner, Member Posts: 295 RM Product Management
    Hello Vasu,
    vasuraghavan wrote:

    I am using RapidMiner SVM learning algorithm for learning from examples. The data is in Excel Spread sheet. If I upload the data to a database table (MySQL) and use DatabaseExampleSource, the process works fine. If I change the DatabaseExampleSource to ExcelExampleSource I get an error "Polinomial attributes not supported. Use Binary2Multiclass learner to wrap the learning algorithm?
    every learner has certain capabilities which means that it can handle certain types of attributes as well as a certain type of label. Generally, SVMs can only handle numerical attributes and a binominal label (i.e. the label is allowed to have only 2 classes). If nominal attributes are present, you may however transform them into numerical ones (either by simply using a distinct number for any nominal value via the [tt]Nominal2Numeric[/tt] operator or by transforming them to binary attributes (each capturing the occurance of a nominal value) via the [tt]Nominal2Binominal[/tt] operator, the latter probably being the better choice. If you have a polynominal label, you may wrap the SVM learner into a Binary2Multiclass meta learner. This transforms a polynominal learning problem into multiple binary classification problems and allows to use the SVM for polynominal learning tasks as well. The [tt]LibSVMLearner[/tt] by the way has a built in multiclass wrapper so that you can use it directly on polynominal classification tasks without wrapping it into a [tt]Binary2MulticlassLearner[/tt].

    Since the process actually worked when the data was loaded from a database, I rather assume your problem is, that the attribute values are not properly presented in the excel file. Might it be that numbers are stored as text in the excel sheet? If that is not the case, feel free to post your process xml file. That makes it easier for us to conclude where your problem actually emerges.

    Hope this was helpful,
    Tobias
  • Legacy UserLegacy User Member Posts: 0 Newbie
    Thank you very much for your reply. Your suggestion was very helpful. My Excel spread sheet had all numbers stored in text format. I changed the format of the cells to numbers and it is working now.

    Please mark this problem as resolved.

    I thank you once again for the help

    Regards

    Vasu Raghavan
Sign In or Register to comment.