Jasmine is testing a machine learning (ML) algorithm that classifies images as being ducks or not ducks. Her model has been built and trained. The testing data consists of 10 images of ducks, and the model correctly identifies them (within an appropriate margin of error) to be ducks. What is wrong with her testing procedure?(1 point)

Question

Jasmine is testing a machine learning (ML) algorithm that classifies images as being ducks or not ducks. Her model has been built and trained. The testing data consists of 10 images of ducks, and the model correctly identifies them (within an appropriate margin of error) to be ducks. What is wrong with her testing procedure?(1 point)
Responses

The testing data is too large.
The testing data is too large.

The testing data is homogenous.
The testing data is homogenous.

The testing data is biased.
The testing data is biased.

The testing data is outdated.

GPT-4o mini · Answer

The most appropriate response is: **The testing data is homogenous.** Since Jasmine's testing data consists solely of 10 images of ducks, it does not provide a diverse range of examples, which is essential for effectively evaluating the performance of the model. A diverse testing dataset should include images of non-ducks and various types of ducks in different conditions and settings to better assess the model's ability to generalize.