classification results of new instances

User43239

I am a novice in using the RapidMiner. My question is "How can I see the classification results of new instances after creating a classifier using a traning dataset?". Many thanks in advance.

rfuentealba

Hello,

Let's go step by step:

- Retrieve your training data. Let's say Titanic Training from the samples dataset on the Repository pane.
- Drag the algorithm you want to train. Let's say Random Forest from the Modeling > Predictive > Trees folder in the Operators pane.
- Connect the out port of your Titanic Training data to the tra port on the Random Forest operator.
- Drag the Apply Model algorithm from the Scoring folder in the operators pane.
- Connect the mod port on the Random Forest operator to the mod port of the Apply Model operator you just dragged.
- Retrieve your unlabeled data. Let's say Titanic Unlabeled from the samples dataset on the Repository pane.
- Connect the out port from the Titanic Unlabeled data to the unl port in the Apply Model operator.
- Connect both ports from the Apply Model operator to the res ports in the end of your canvas.

If you do this, you may be able to see this:



If you press play, the Apply Model will just send results to the Results tab. What you will see is a model (or a representation of it), and the results of your new instances after creating a classifier using a training dataset. Just like this:



The prediction is your target. Tweak a few times and come back to execute. This is your window.

BTW, RapidMiner Academy has all this knowledge available. You may want to check it out.

All the best,

Rodrigo.