Find the lowest value of an attribute in relation to the other

onep

Hello.

I'm just getting started to learn Radpidminer and would like some help to my first project.

I've managed to import my data into RM and made a predictive process on some training data and can predict my unlabeled data based on the trained data.

Now, I've trained my model with the raw training data, but I had realized that I need to cleanse the data first. See I only want my model to be trained with the data with the lowest possible value of the column B.

 

So here is my setup;

Column A - Labeled data

Column B - Data where i only want the lowest value

Coumn C - Data

 

Example:

First row

Column A: 100

Column B: 20

Column C: 50

Second row

Column A: 100

Column B: 19

Column C: 50

 

Now, since A+C is matching I would only want my model to be trained with the second row, since the first row has a higher value of column B.

 

Is this achiveable in Rapidminer?

I do not hope that my explanation is too confusing 

 

 

onep

I only want to maintain the lowest value if the other columns are the same. If I invent an example with a car, it might make more sense .

I have the following data from the car in my database;

Speed, gear selection, fuel consumption.

 

I've got alot of rows with data from the car, for example;

100 km / h, 5th gear, 0.3 l / km

100 km / h, 4th gear, 0.35 km / l

 

Gear selection is my label. In this case, I keep only the row with the 5th gear because it gives me the lowest fuel consumption.

 

But if I have a row with the following data;

95 km / h, 4th gear, 0.35 km / l

I would still keep this row, becuase now speed has changed.