Free AWS MLA-C01 Actual Exam Questions - Question 2 Discussion
An ML engineer is developing a fraud detection model on AWS. The training dataset includes transaction logs, customer profiles, and tables from an on-premises MySQL database. The transaction logs and customer profiles are stored in Amazon S3. The dataset has a class imbalance that affects the learning of the model's algorithm. Additionally, many of the features have interdependencies. The algorithm is not capturing all the desired underlying patterns in the data. The ML engineer needs to use an Amazon SageMaker built-in algorithm to train the model. Which algorithm should the ML engineer use to meet this requirement?
Guessing A, LightGBM handles imbalanced data and complex feature interactions nicely.
Good point about clustering and topic modeling being off for fraud detection, so A it is.
Makes sense to rule out C and D right away since clustering and topic modeling don’t fit a supervised fraud detection task. Between A and B, I’d pick A too because LightGBM is specifically designed to handle class imbalance and complex feature relationships, which linear learner might struggle with. Plus, it’s known for better accuracy in these scenarios.
It’s A; LightGBM’s gradient boosting is great for class imbalance and complex features.
B seems limited for feature interactions here, right? LightGBM (A) is better for that.
Maybe A, LightGBM handles class imbalance and feature interactions better.
I think it’s safe to rule out C and D since this is clearly a supervised learning problem, not clustering or topic modeling. Between A and B, the class imbalance and feature interdependencies point more toward something that can handle complex interactions, so LightGBM (A) seems stronger. Also, Linear Learner (B) might struggle with non-linear patterns here. But I’m curious if the integration of on-prem MySQL data impacts which algorithm SageMaker’s built-ins can handle efficiently? Would that change the choice?
Makes sense to skip C and D since this isn’t unsupervised or topic modeling. Between A and B, LightGBM (A) better handles complex patterns and imbalance, which fits the fraud detection need here. A
Option A seems right since LightGBM can handle class imbalance and feature interactions well, unlike Linear Learner which might struggle with complex dependencies. Clustering and NTM don’t fit supervised fraud detection tasks.
Maybe A? LightGBM handles class imbalance better and can capture feature interactions, unlike linear or clustering algorithms. Not sure about NTM fitting this use case though.