"Selecting attribut with high weight from producd weight file"

shk721

Hi;


i am mow looking for method to sort the attribute generated by weightining process. The aim is to select percentage of the data based on the order.


Actually , i have triyed to edit xml which is the format of this file .However , is there any simple way to do that.


Regard;
Hassan

steffen

Hello Hassan

Since I am not sure whether you want to select a subset of features  or a subset of rows, I will explain both.

Subset of features
Suppose you have already generated AttributeWeights.
I suggest in Preprocessing->Attributes->Selection the operator "AttributeWeightSelection". This operator allows you to select a subset of features based on the AttributeWeights. Take a look at the tutorial.pdf (available at the rapidminer download page) where you will find a more detailed explanation of this operator.

Select a subset of rows
Suppose you got an attribute named "weight" which is already a part of the ExampleSet. I suggest in Preprocessing->Data->Filter the operator "ExampleFilter". For Example with parameters like this:

 <operator name="ExampleFilter" class="ExampleFilter">
        <parameter key="condition_class"	value="attribute_value_filter"/>
        <parameter key="parameter_string"	value="weight&gt;=0.5"/>
    </operator>

This will return only the examples whose value of the attribute "weight" is greater than 0.5.
Take a look at the tutorial.pdf (available at the rapidminer download page) where you will find a more detailed explanation of this operator.

hope this was helpful

Steffen