AWS Certified Machine Learning – Specialty — Question 266

A machine learning (ML) specialist is training a multilayer perceptron (MLP) on a dataset with multiple classes. The target class of interest is unique compared to the other classes in the dataset, but it does not achieve an acceptable recall metric. The ML specialist varies the number and size of the MLP's hidden layers, but the results do not improve significantly.

Which solution will improve recall in the LEAST amount of time?

Answer options

• A. Add class weights to the MLP's loss function, and then retrain.
• B. Gather more data by using Amazon Mechanical Turk, and then retrain.
• C. Train a k-means algorithm instead of an MLP.
• D. Train an anomaly detection model instead of an MLP.

Correct answer: A

Explanation

Adding class weights to the loss function is the fastest solution because it penalizes misclassifications of the minority target class, directly improving recall without requiring architectural changes or new data collection. Gathering more data via Amazon Mechanical Turk is time-consuming and expensive. Switching to k-means (unsupervised clustering) or anomaly detection does not leverage the existing multi-class labels effectively for this supervised classification problem.