Correlation Attributes with SVM

theopilus27

im sorry my english very bad.
ok. i am implementation SVM for major in higher students school
for my case with two classification (IPA & IPS) with 17 attributes and two attributes prediction (IPA& IPS)
this is my example dataset :

A      B    C  D    E    F    G
0.1  0.2  0.3 9.0  8.0 7.0  IPA

my questions is :
1) how i can implementation min-max normalization with rapidminer ?
2). how i can find correlation attributes ? i have been tried using Correlation matrix but can unable pairwise attribute (example : attributes A & B correlation 0.7) I would expect to show 3 or more correlation attributes..

Please help me..

Thx u very much

MariusHelf

Hi,

1) You can use the Normalize operator.
2) The Correlation Matrix operator has 3 outputs. Did you have a look at all of them?

Best regards,
Marius

theopilus27

Thx u very much Marius..
has 3 outputs?? mm, in my display unable pairwise attribute (example : attributes A & B correlation 0.7)..why different??
i used 10 fold cross validation. if after i  was used 10 fold cross validation, what i do need to do divided my data set into two parts
( training and testing) ??
"because my literature said that the data should divided into two parts ( training and testing) with 50%-50%. testing data is used to labeling.."

MariusHelf

Yes, the Correlation Matrix has three outputs. Just connect the second one called "mat" to the process output and have a look at the Pairwise Table - that's what you are looking for.

Regarding the cross validation: if you use a cross validation, you don't need a so-called hold-out validation in addition. Just google for the cross validation to find out about its concepts and its use. There is a lot of material about it on the web.

Best regards,
Marius

theopilus27

Thx u very much Marius..very interesting for me...
my next questing is if i am using normalization (for example minmax and decimal point normalization) what this change structur my dataset ??
and
how do i know processing time consume when classification in rapidminer ???

theopilus27

and also about Min-max Normalization with up (1) and bottom (-1) with Normalize-operator,
i ve tried it but why the result is different with my calculate with excel..??

<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<process version="5.3.008">
  <context>
    <input/>
    <output/>
    <macros/>
  </context>
  <operator activated="true" class="process" compatibility="5.3.008" expanded="true" name="Process">
    <process expanded="true">
      <operator activated="true" class="retrieve" compatibility="5.3.008" expanded="true" height="60" name="Retrieve svmthesis_original" width="90" x="112" y="30">
        <parameter key="repository_entry" value="//Local Repository/svmthesis_original"/>
      </operator>
      <operator activated="true" class="normalize" compatibility="5.3.008" expanded="true" height="94" name="Normalize" width="90" x="246" y="30"/>
      <connect from_op="Retrieve svmthesis_original" from_port="output" to_op="Normalize" to_port="example set input"/>
      <connect from_op="Normalize" from_port="example set output" to_port="result 1"/>
      <portSpacing port="source_input 1" spacing="0"/>
      <portSpacing port="sink_result 1" spacing="0"/>
      <portSpacing port="sink_result 2" spacing="0"/>
    </process>
  </operator>
</process>

my formula in excel is

=(B3-B$291)/(B$292-B$291)*(1-(-1))+1

Please help me sir..thx u..

MariusHelf

Hi,

you have to enable expert mode in RapidMiner (if not yet done so), and set the method parameter of the Normalization operator to range_transformation.

Best regards,
Marius