About the LinearRegression operator

Legacy User

Hi,

I am a new user of rapidminer. I tried to use the LinearRegression operator and I ran with my dataset. But in the end I did not get an equation. I am currently trying to train a model.

My process are as follows:

Root

 >CSV Example Source

 >Attribute Filter

 >RemoveUselessAttributes

 >Genetic Algorithm

   >>Operator Chain

      >>>X-Validation

            >>>>LinearRegression

            >>>>OperatorChain(2)

                     >>>>>>ModelApplier

                     >>>>>>RegressionPerformance

            >>>>ProcessLog

 >CSV Example Set Writer


Thanks!

haddock

Hi there,

The optimisation produces a parameter set for the learning operator and anything else you tweak, rather than a model. To keep it simple the process is like this.

1. Generate a parameter set - you've done that.

2. Apply that set using the parameter setter operator. You'll need to map between the optimised parameters and the ones you will use. So if I had optimised an SVM and some Validation settings I would have this sort of mapping.

<operator name="ParameterSetter" class="ParameterSetter" activated="no">
            <list key="name_map">
              <parameter key="NNValidation" value="NNValidation (2)"/>
              <parameter key="LIbSVMLearner" value="LibSVMLearner (2)"/>
            </list>
        </operator>

3. Run the same learner on an appropriate dataset ( careful here ). That will produce the optimised model which has this sort of form.

31.736 * a1 + 42.948 * a2 + 23.773 * a3 + 3.706 * a4 - 4.184 * a5 - 304.228

Where a1 etc. are the attribute names.

With RM a big skill is to keep track of the inputs and ouputs of the operators, because sometimes it is not what you think!

Legacy User

Hi haddock,

Thanks for your suggestion.

So the parameter setter operator should include two learner operators and placed after genetic algorithm?

Thanks

haddock

No, just the one learner after setting the parameters.