Free Google Professional-Machine-Learning-Engineer Actual Exam Questions - Question 13 Discussion
sources and sends them to users. You need to build a recommendation model that will suggest
articles to readers that are similar to the articles they are currently reading. Which approach should
you use?
B. Since the goal is to find articles similar to the one currently being read, using vector embeddings and similarity makes the most sense. Collaborative filtering (A) depends on user history, which might not help when a user reads new topics. Logistic regression for each user (C) seems impractical given the variety of users and articles. Manually labeling articles (D) is too time-consuming and focuses on categories, not similarity, so B still fits best here.
It’s B, since vector similarity directly measures article closeness without needing user data.
B imo, because vector similarity works better for content-based rec than categories.
B vs D; D focuses on categories, not article similarity like B does.
A imo, since collaborative filtering isn’t great without detailed user history.
B, since collaborative filtering (A) needs strong user history, which isn’t mentioned here.
It’s B because content-based similarity fits better than user-focused methods here.
Guessing B, since it directly matches articles by content similarity, not user history.
B since it focuses on article content, matching the question’s similarity requirement directly.
Maybe B makes the most sense here since the goal is to find articles similar to what the user is currently reading. Collaborative filtering (A) depends heavily on user behavior data, which isn’t mentioned. Also, C sounds like overkill and would require tons of user-specific models. D requires manual labels, which isn’t practical for large, constantly updating news feeds. Encoding articles as vectors with word2vec and comparing similarity seems like the most straightforward way to capture content similarity directly.
Totally agree, B is best for content similarity without needing user data. B
Option B makes the most sense since it directly measures article content similarity, which is what the question asks for. Collaborative filtering (A) relies on user data, which isn't mentioned here.
I think B is the way to go since it directly captures article content similarity using vector embeddings, which aligns perfectly with recommending similar articles rather than user behavior.
B This seems the best fit since it directly compares article content via vector similarity, which matches the goal of finding articles like the one currently being read, unlike user-focused methods.
Makes sense to me that A is off because it focuses on user history, not article similarity itself. The question wants articles like the current one, so B fits better. A
A imo, because collaborative filtering relies on user behavior patterns, not the content similarity of articles. Since the question wants articles similar to the one currently being read—not just what users liked before—B seems smarter for content-based recommendations. C’s per-user logistic regression feels way overcomplicated, and D’s manual labeling is impractical here. So B is the best fit since vector similarity directly compares article content, not user preferences or categories.
It’s B, since vector similarity is perfect for finding similar articles quickly.
B/C? B makes sense since vector similarity fits the “similar articles” goal, while C feels too personalized and complex here. A’s more for user preference, not article similarity. D’s way too manual and slow.