Q: 8
You have been tasked with creating a model that will recommend products based on what other customers have similarly purchased. Which algorithm is the best choice given this situation?
Options
Discussion
A tbh. Most recommenders use KNN/collaborative filtering for exactly this, but if it had talked about predicting intent or super complex user vectors, C could possibly edge in. Here, classic approach wins unless "best" means something unusual.
Its A, K Nearest Neighbor is classic for collaborative filtering tasks like these. Neural nets are overkill here. Anyone disagree?
Option C. had something like this in a mock. Neural networks work well for recommendations.
A is my pick. KNN is basically made for collaborative filtering like this, matching customers with similar purchase behavior. Unless the question asks about massive scale or deep personalization, pretty sure A works best.
Option A K-means (B) is a classic trap here, but it's not a recommender method. KNN fits the scenario best.
Seen this kind of question in some practice sets, almost always goes with A. KNN is standard for finding users/items with similar purchase history, which fits collaborative filtering. Not 100% but all signs point to A.
Maybe A, but only if the question expects recommendations based solely on direct purchase similarity-if they'd asked for segment-based, B might've fit.
I might lean toward B here. K-means could help cluster similar customers, so you'd recommend products within the same group. I'm not fully sold though since the question asks for direct recommendation based on customer similarity, which maybe fits A better. But grouping first feels logical to me in practice. Anyone disagree?
Option A Had something like this in a mock, pretty sure KNN is what they want for these collaborative-type recommendations.
B , K-means is clustering so not really what you want unless you're grouping customers first. But if the question was phrased around segment discovery before recommendations, B might make sense. Otherwise, A seems stronger but not 100% sure.
Be respectful. No spam.