Precision

Part 5 of Performance Metrics for Machine Learning Models (Classification Metrics) by Dr. Alvin Ang

This is part of a series on :

Dr. Alvin Ang

Performance Metrics for Machine Learning Models in Python

View list

13 stories

When to use What?

Key Purpose of Obtaining Precision

Objective for getting both Precision and Recall is to obtain the F1 score (as we will see later).

What is Precision?

Precision is between 0 to 1

• 0 = BAD
• 1 = GOOD

So if we have Precision = 0.9… we can say…

This Model is very Precise because most of its prediction on Positives are Really Actually True…

Precision in Python

Precision Score on PIMA Indian Diabetes Data

Logistic Regression on Pima Diabetes Dataset

The model can Precisely predict

Precision for Class 0 = 83%.

Out of all instances predicted as 0, what percentage were actually 0.