Precision and Recall – Classification Metrics

What is the definition of precision in a confusion matrix?

What is the formula for precision?

What is the definition of recall in a confusion matrix?

What is the formula for recall?

When should you use precision and recall as your performance metric for your model?

Which answer is incorrect about using a performance metric for a model?

10 people were suspected of having coronavirus. 3 people were found infected and 7 were not infected after laboratory reports. Our model predicts that 8 people were infected, and 2 people were not infected. What is the value of Precision?

A Data Scientist at Inabia has fine-tuned a pre-trained model with Yolov5 and his own custom dataset. There are 15 images in the test dataset including 200 people in total. Which option is correct on calculating the model’s precision and recall based on the confusion matrix as shown below (Positive: no face mask, Negative: face mask):

You are building a face mask detection system at a bus station and evaluating the model’s performance with a confusion matrix (see the table below). Detecting the people who are not wearing face masks is the most important target in this case. Which one is the best metric in the scenario? Calculate the number of the metric in this case.

A company is designing a face mask alarm at the entry of a grocery store. The detector will send a warning if it predicts a person who is passing the entry without wearing a face mask. The company would like to evaluate the model performance based on the collected data with a confusion matrix (see the table below). The criteria of Recall should be equal or over 99.5%. According to the metric, does the model meet the requirement of performance and what is the current number of the recall?