Neural network can not handle binominal attributes

siamak_wantsiamak_want Member Posts: 98 Contributor II
edited November 2018 in Help
Hi RM's fans,

I am about to use neural network on a dataset witch all of the attributes are binominal (boolean). I got the message that the neural network operartor can not handle binominal attributes. So I put a NominalToNumerical attribute to solve the problem. But I am not sure I did the right way. I doubt dealing with boolean attributes as neumerical may cause bad effects in my analysis.

Please inform me if the way I exploited is right. Any idea would be appreciated.

best regards,


  • Options
    TheBenTheBen Member Posts: 11 Contributor II
    try to use Meta Modelling or another model
  • Options
    wesselwessel Member Posts: 537 Maven
    It is not possible to answer this question without additional information.
    My guess would be, encode true as 1, false as -1. And make sure you use tangent nodes, not sigmoid.

    But please make sure you read some literature so that you know what you are doing.
    Neural networks are really well documented.

    This I got from Rapid Miner documentation (F1)
    The used activation function is the usual sigmoid function. Therefore, the values ranges of the attributes should be scaled to -1 and +1. This is also done by this operator if not specified otherwise by the corresponding parameter setting. The type of the output node is sigmoid if the learning data describes a classification task and linear for numerical regression tasks.

    Please also check out W-MultiLayerPerceptron, this one is from Weka.

    You may like this paper, it uses also some binary attributes.
Sign In or Register to comment.