Amazon MLS-C01 Exam - Topic 1 Question 123 Discussion

Actual exam question for Amazon's MLS-C01 exam

Question #: 123
Topic #: 1

[Exploratory Data Analysis]

A Machine Learning Specialist is deciding between building a naive Bayesian model or a full Bayesian network for a classification problem. The Specialist computes the Pearson correlation coefficients between each feature and finds that their absolute values range between 0.1 to 0.95.

Which model describes the underlying data in this situation?

AA naive Bayesian model, since the features are all conditionally independent.

BA full Bayesian network, since the features are all conditionally independent.

CA naive Bayesian model, since some of the features are statistically dependent.

DA full Bayesian network, since some of the features are statistically dependent.

Show Suggested Answer

Suggested Answer: D

by Lettie at Oct 30, 2025, 02:09 AM

Limited Time Offer

25%

Off

Get Premium MLS-C01 Questions as Interactive Web-Based Practice Test or PDF

Contribute your Thoughts:

Submit Cancel

Chanel

9 days ago

I agree with D too. Full Bayesian networks handle dependencies well.

upvoted 0 times

...

Xenia

14 days ago

I think D is the best choice. Some features are dependent.

upvoted 0 times

...

Leslee

19 days ago

Naive models are great, but with those correlations, D makes more sense.

upvoted 0 times

...

Lilli

24 days ago

Pearson values show some correlation, so C could be valid too.

upvoted 0 times

...

Gregg

30 days ago

Wait, are we sure about the dependencies? Seems tricky.

upvoted 0 times

...

Samira

1 month ago

Totally agree, definitely leaning towards D.

upvoted 0 times

...

Edna

1 month ago

I'm with Clay on this one. The Pearson coefficients show that the features aren't all independent, so a full Bayesian network is overkill. Gotta keep it simple, folks.

upvoted 0 times

...

Clay

2 months ago

Option C is the way to go. With some features being statistically dependent, a naive Bayesian model is the way to go. Simple is sometimes better, you know?

upvoted 0 times

...

Shanda

2 months ago

Haha, this is a classic case of "it depends." The question doesn't give us enough info to say for sure. Maybe the ML Specialist should just flip a coin and call it a day.

upvoted 0 times

...

Aleta

2 months ago

Hmm, I'm not so sure. If the features are all conditionally independent, then a full Bayesian network might actually be the better choice. Gotta think this through carefully.

upvoted 0 times

...

Yun

2 months ago

I think I read that if there are dependencies, we should go with a full Bayesian network. So, I’m leaning towards option D, but I’m not completely confident.

upvoted 0 times

...

Kirk

2 months ago

I’m a bit confused. The high correlation could mean we need a full Bayesian network, but I thought naive Bayes could still work if we treat them as independent.

upvoted 0 times

...

Osvaldo

2 months ago

Looks like option D is the way to go. Features are dependent!

upvoted 0 times

...

Genevieve

3 months ago

This question feels similar to one we practiced where we had to decide based on feature independence. I think if some features are dependent, it leans towards a full Bayesian network.

upvoted 0 times

...

Judy

3 months ago

A naive Bayesian model would be the way to go here. Those correlation coefficients are all over the place, so the features can't be conditionally independent.

upvoted 0 times

...

Sherron

3 months ago

I remember that naive Bayes assumes conditional independence, but the correlation coefficients suggest some dependencies. I'm not sure which model fits better.

upvoted 0 times

...

Dottie

4 months ago

I think the key here is to focus on the Pearson correlation coefficients. Since they range from 0.1 to 0.95, that means some features are dependent, so a full Bayesian network would be the appropriate model to use.

upvoted 0 times

...

Salley

4 months ago

I'm pretty confident that the answer is D. A full Bayesian network should be used since the features have varying degrees of correlation, indicating they are not all conditionally independent.

upvoted 0 times

...

Bobbye

4 months ago

Okay, let me think this through. If some of the features are dependent, then a full Bayesian network would be the better choice to model those relationships. The key is that the question mentions the Pearson correlations, which indicate statistical dependence between the features.

upvoted 0 times

...

Rory

4 months ago

I'm a bit confused here. If the features are conditionally independent, then a naive Bayesian model should work, right? But the question says some of the features are dependent, so I'm not sure which model to choose.

upvoted 0 times

...

Farrah

4 months ago

Hmm, the Pearson correlation coefficients range from 0.1 to 0.95, so that suggests some of the features are dependent. I think a full Bayesian network might be better to capture those dependencies.

upvoted 0 times

Lynette

3 months ago

I agree, the high correlation indicates dependencies.

upvoted 0 times

...

Adolph

3 months ago

Exactly, we need to model those dependencies accurately.

upvoted 0 times

...