Read CSV

SamSalas

HI  

 

I'm trying read a CSV file with missing values and apply a model. 

My problem is that any missing values in unlabelled data change his type and a error appear "The imput ExampleSet dpes not match the training ExampleSet..." I cant change the types manually beacause i have more than 4000 attributes.

I tryied  use Nominal to Numerical and I lose all missing datas with the convesion.

 

How i force all attributes to be numerical when i import it? Thanks for your answers.

 

 

Thomas_Ott

The error alludes to a different problem. What it means is that what ever the name, quantity, and type of the attribute columns you used to train the Neural Net model on, the testing set does not match that. Simplistically speaking, if you trained the model on 10 attributes and test it on 9 or 11, it will not work. 

 

If you want to force numericals, you can do it one by one in the Read CSV operator OR use Guess Types and/or Parse Numbers operator.