Confusion regarding precision vs weighted precision

shashwat01
shashwat01 Altair Community Member
edited November 5 in Community Q&A

The question is related to the project. In connection with Performance Metrics - Classification. I'm struggling to identify the metrics name in rapidminer list of performance metrics. Whereas in the course material we defined accuracy, Precision, Recall and F1 Score. In rapidminer the list of output in Performace only has a match with Accuracy. Can you help me identifying which Rapidminer metrics relate to Precision, Recall and F1 Score?


Tagged:

Best Answer

  • RolandJones
    RolandJones
    Altair Employee
    Answer ✓
    Hi @shashwat001,

    Precision, Recall, and F1 Score are metrics for Binominal Classification - i.e. you can only have one of two possible answers. In Performance (Classification), we expose the performance metrics for classification problems with more than two possible answers. For Precision and Recall, I would recommend looking at the Weighted Mean Recall/Precision options. I don't believe we've enabled the F-measure metric for multi-class problems, but it should be fairly easy to calculate using the formula if it's needed.

    Best,

    Roland

Answers

  • RolandJones
    RolandJones
    Altair Employee
    Answer ✓
    Hi @shashwat001,

    Precision, Recall, and F1 Score are metrics for Binominal Classification - i.e. you can only have one of two possible answers. In Performance (Classification), we expose the performance metrics for classification problems with more than two possible answers. For Precision and Recall, I would recommend looking at the Weighted Mean Recall/Precision options. I don't believe we've enabled the F-measure metric for multi-class problems, but it should be fairly easy to calculate using the formula if it's needed.

    Best,

    Roland

Categories