AWS Certified Machine Learning Specialty · Free Practice Question Medium
Question 5
An insurance company is building a binary classification model to predict insurance claims. The training data contains 1800 instances of the positive class (customers who did claim insurance) and 100 instances of the negative class (customers who did not claim insurance). The final model has 85% accuracy, but poor precision.
How can you improve the model performance? (Select three)
-
A
Create more samples using algorithms such as SMOTE
-
B
Collect more training data for the negative class
-
C
Over-sample from the negative class
-
D
Over-sample from the positive class
-
E
Collect more training data for the positive class
Reveal correct answers
Correct answers: A, B, C
Explanation
Correct options:
Over-sample from the negative class
Collect more training data for the negative class
Create more samples using algorithms such as SMOTE
In case of a binary classification model with strongly unbalanced classes, we need to over-sample from the minority class, collect more training data for the minority class and create more samples using algorithms such as SMOTE which effectively uses a k-nearest neighbours approach to exclude members of the majority class while in a similar way creating synthetic examples of a minority class. Here are a few good references :
http://www.svds.com/learning-imbalanced-classes/
Incorrect options:
Over-sample from the positive class
Collect more training data for the positive class
Over-sampling from the positive class or collecting more training data for the positive class would further worsen the model.
Discussion
Think the marked answer is wrong, or have a better explanation? Share it below — comments appear after review.
