Classifier Models > Classifier Scores
  

Classifier Scores

A Classifier transformation compares each row of input data with every row of reference data in a classifier model. The transformation calculates a score for each comparison. The scores represent the degrees of similarity between the input row and the reference data rows.
When you run a mapping that contains a Classifier transformation, the mapping returns the label that identifies the reference data row with the highest score. The score range is 0 through 1. A high score indicates a strong match between the input data and the model data.
Review the classifier scores to verify that the label output accurately describes each row of input data. You can also review the scores to verify that the classifier model is appropriate to the input data. If the transformation output contains a large percentage of low scores, the classifier model might be inappropriate. To improve the comparisons, compile the model again. If the compiled model does not improve the scores, replace the model in the transformation.