"Newbie: clustering with k-means and then using multipe regression linear"

abudcoolabiz

Hi, I'm a student from Indonesia.

Currently I'm working on my assignment to predict lead time manufacturing on make to order manufacture. The dependent attribute is lead time (continuous). and the independent variable are routing, material, etc.

I'm going to use clustering (k-means) to make a few groups of data. and then I will make rules of each group using multiple regression linear.

I'm very new to this software. So the first question is, Can I do these method with Rapid Miner? Then how can I accomplished it? try with simple and detail step please... thank you so much for the attention. and I really hope you can help me.

best regards,

land

Hi,
for getting more familiar with the software, I really recommend to do the online tutorial available in the welcome screen.
If you have a basic understanding of the software, here is it, how it works for your assignment. But because I'm not getting paid for solving your exercises, I must hold this quite short, but you will get an idea how to work this out.

Here's a process setup, which will first load data (in this case it will be generated randomly). Then KMeans is used as clustering algorithm, which will assign each example to a cluster. This cluster attribute is then used as label, hence the target variable for a following linear regression. You might exchange operator with an analog operator, for example use LinearDiscriminantAnalysis instead of linear regression, or use another clustering operator.

<operator name="Root" class="Process" expanded="yes">
    <operator name="ExampleSetGenerator" class="ExampleSetGenerator">
        <parameter key="target_function"	value="spiral cluster"/>
        <parameter key="number_of_attributes"	value="2"/>
    </operator>
    <operator name="KMeans" class="KMeans">
    </operator>
    <operator name="ChangeAttributeRole" class="ChangeAttributeRole">
        <parameter key="name"	value="cluster"/>
        <parameter key="target_role"	value="label"/>
    </operator>
    <operator name="LinearRegression" class="LinearRegression">
    </operator>
</operator>

I hope that this will help you, solving your problem.

Greetings,
  Sebastian