Precision Metric

A Precision Metric is a performance metric for a binary classification model that is based on the probability that a positive prediction is a correct prediction.

• AKA: Precision, Positive Predictive Value, Precision Rate, Binary Classifier Precision Metric.
• Context:
  • It can be estimated by: TP/(TP+FP)
  • It can be negatively correlated with a recall metric (at the expense of).
• Example:
  • # of correct answers given by the system as a proportion of the total # of answers given.
  • Set Intersection(Relevant Artifacts, Retrieved Artifacts) / Retrieved Artifacts.
• Counter-Example(s):
  • a Recall Metric.
  • an F-Measure.
  • an Accuracy Metric - the probability that a prediction is a correct prediction.
• See Also: True Positive Rate, Fallout, Numeric Attribute Precision, Reproducibility Metric.

References

2010

• http://en.wiktionary.org/wiki/precision#Noun
  • 1. the state of being precise or exact; exactness
• http://en.wikipedia.org/wiki/Precision_and_recall
  • Precision and recall are two widely used statistical classifications.
  • Precision can be seen as a measure of exactness or fidelity, whereas Recall is a measure of completeness.

2000

• 2000 SpeechAndLanguageProcessing
  • "Precision is a meaure of how much of the information that the system returned is actually correct."

1998

• (Kohavi & Provost, 1998) => Ron Kohavi, and Foster Provost. (1998). "Glossary of Terms." In: Machine Leanring 30(2-3).
  • Precision: d/(b+d).