引導式測驗 · 中等 等級

協同濾波測驗

測試使用者項目回饋、類似讀者推薦、顯性訊號與隱性訊號、潛在因素、冷啟動、暴露偏差和基線。

相關引導路徑協同過濾
問題 1 的 8

What information is central to collaborative filtering in this guide?

預覽每個問題8 問題

當您參加測驗時,就會出現答案和解釋。

  1. What information is central to collaborative filtering in this guide?
    • Patterns in how users interact with items
    • Only a writer's hand-made genre label for each item
    • Only the physical size of an item
    • A guaranteed statement of each user's future preferences
  2. A and B both liked books P and Q; B also liked R, which A has not seen. Why might R be suggested to A?
    • R must share the same author as P and Q
    • A's missing rating for R proves dislike
    • The two shared books guarantee that A will enjoy R
    • Their overlapping histories make B's interest in R a candidate signal
  3. Why should a brief video play be treated differently from an explicit positive rating?
    • A play may reflect exposure or curiosity rather than satisfaction
    • Implicit activity always expresses a stronger preference
    • Explicit ratings are never part of recommender data
    • A play reveals the user's full reason for watching
  4. What does matrix factorization learn from a user–item feedback matrix?
    • A legal identity record for each user
    • Compact user and item vectors whose relationship predicts interactions
    • Human-labeled genres that every latent dimension must match
    • Only one popularity ranking with no user differences
  5. Why is a brand-new item difficult for a purely interaction-based recommender?
    • Its content features are necessarily illegal to use
    • The user–item table always contains its future sales
    • Every new item has already been rated negatively
    • It has little or no interaction history from which to learn
  6. An item was never shown to a user. What should the system avoid inferring from the missing click?
    • That exposure may shape what can be clicked
    • That the feedback table has an unobserved entry
    • That the user definitely disliked the item
    • That a later recommendation could still be tested
  7. How can recommending only popular items create a feedback loop?
    • Popularity automatically proves every item is higher quality
    • A popular item stops collecting any feedback
    • Promoted items gain more exposure and interactions while less-shown items remain sparse
    • All users receive equally diverse discovery
  8. Which comparison helps test whether a collaborative model adds value?
    • Only the model's embedding dimension
    • Time-aware results against popularity and content-based baselines, with user outcomes
    • Only its score on interactions from the training period
    • A claim that latent factors are always better than item descriptions