IOMultiplier and IOSelector and other basic questions

Legacy User

hi all,

I am new to RapidMiner and have some basic questions now:

1) IOMultiplier and IOSelector

can somebody give me a full working example of how to use IOMultiplier? what I want to do is to duplicate a dataset to create two separate operator chains for preprocessing the data.

the following micky mouse example is working, but validating the process structure throws an error ([Error] Mapping_1: Mapping_1: Mapping_1: Missing input: ExampleSet).

<operator name="Root" class="Process" expanded="yes">
    <operator name="DatabaseExampleSource" class="DatabaseExampleSource">
        <parameter key="database_system"	value="PostgreSQL"/>
        <!-- connection settings -->
    </operator>
    <operator name="IOMultiplier" class="IOMultiplier">
        <parameter key="io_object"	value="ExampleSet"/>
        <parameter key="number_of_copies"	value="2"/>
    </operator>
    <operator name="IOSelector_1" class="IOSelector">
        <parameter key="io_object"	value="ExampleSet"/>
		<parameter key="select_which"	value="1"/>
    </operator>
    <operator name="Mapping_1" class="Mapping">
        <parameter key="add_default_mapping"	value="true"/>
        <parameter key="attributes"	value="attribute_name"/>
        <parameter key="default_value"	value="17"/>
    </operator>
    <operator name="IOSelector_2" class="IOSelector">
        <parameter key="io_object"	value="ExampleSet"/>
        <parameter key="select_which"	value="2"/>
    </operator>
    <operator name="Mapping_2" class="Mapping">
        <parameter key="add_default_mapping"	value="true"/>
        <parameter key="attributes"	value="attribute_name"/>
        <parameter key="default_value"	value="4"/>
    </operator>
</operator>

is there a way to get a graphic representation of separate operator chains in the tree structure of the operator tree view? (with n child nodes for an IOMultiplier)


2) change attribute type

how can I change the attribute type from integer to nominal without defining a mapping? what I would like to have is something like String.valueOf(int i) applied to each value of a column in my dataset. again, could someone give me a complete working example please?

thanks!

steffen

Hello

the following micky mouse example is working, but validating the process structure throws an error ([Error] Mapping_1: Mapping_1: Mapping_1: Missing input: ExampleSet).

As far as I know, the validation procedure has problems with the count of IOObjects. If your process is working, feel free to ignore the messages. Beside I see no problem with your setup.

is there a way to get a graphic representation of separate operator chains in the tree structure of the operator tree view? (with n child nodes for an IOMultiplier)

Sorry, I dont understand.

how can I change the attribute type from integer to nominal without defining a mapping? what I would like to have is something like String.valueOf(int i) applied to each value of a column in my dataset. again, could someone give me a complete working example please?

Try the operator "Numerical2Polynominal"

regards,

Steffen

Legacy User

hello steffen,

thanks for your reply!

steffen wrote:
Sorry, I dont understand.

i would like to have a graphical representation like this (kind of what it is like in SPSS Clementine):

----------------
| IOMultiplier |
----------------
       |--- stream 0 ---| Operator |----> result 0
       |
       |----stream 1 ---| Operator |----> result 1

     
is this possible with RapidMiner?

Try the operator "Numerical2Polynominal"

as far as i understand it Numerical2Polynominal is applied to every attribute, right?
i want the conversion to be applied to the values of only one attribute (one columnn in my dataset), how can i do this?

thanks!

steffen

Hello

i would like to have a graphical representation like this (kind of what it is like in SPSS Clementine):

I am still not sure if I got you right, but: If you click the check button you get a text-based representation in the log-output below. Another option is View->BoxView.

as far as i understand it Numerical2Polynominal is applied to every attribute, right? i want the conversion to be applied to the values of only one attribute (one columnn in my dataset), how can i do this?

Yes, you are correct. Use the Operator "AttributeSubsetPreprocessing"

regards,

Steffen