An online bookstore wants to increase sales by recommending additional books to customers based on what similar customers have purchased. The company has transaction data from millions of customers but doesn't know which books tend to be purchased together. Which approach would best accomplish this goal?

Question

BlackTom AI · Accepted Answer

Collaborative filtering identifies customers with similar purchase histories and recommends items those similar customers bought but the target customer has not — exactly the capability needed here.

1) Relevant knowledge points and concepts
- Recommender systems: personalize suggestions for users based on data. Two main approaches: collaborative filtering (CF) and content-based filtering.
- Collaborative filtering: finds patterns across users (user-based) or items (item-based) by comparing purchase histories; works well with large transaction datasets.
- Alternative tools: OLAP cubes (exploratory aggregation), data warehouses (storage/standardization), decision support systems (what-if modeling), and descriptive analytics (summaries).

2) Step-by-step reasoning process
- Goal: recommend additional books based on what similar customers purchased.
- Input: millions of transaction records (sufficient data to learn customer similarity).
- Method selection: use data mining techniques for recommendation; collaborative filtering directly models similarity between customers or between items and yields personalized recommendations.

3) Mini calculation/example (intermediate steps)
- Example: Customer A bought {Book1, Book2}; Customer B bought {Book1, Book2, Book3}.
- Jaccard similarity(A,B) = |{1,2} ∩ {1,2,3}| / |{1,2} ∪ {1,2,3}| = 2/3 ≈ 0.67 (high similarity).
- Since B has Book3 and A does not, CF would recommend Book3 to A.
- At scale, matrix-factorization or k-NN over user-item matrices produces ranked recommendations.

4) Why the correct answer is correct
- The option “Use data mining with collaborative filtering…” explicitly performs pattern discovery across similar customers and scales to millions of transactions, producing targeted cross-sell recommendations — matching the company’s objective.

5) Why other options are incorrect
- OLAP cube: useful for managers to slice data, not for producing personalized recommendations automatically.
- Data warehouse: important for storage and consistency, but by itself doesn’t produce recommendations.
- Decision support system for pricing: models pricing scenarios, unrelated to recommending additional books.
- Descriptive analytics: gives aggregate summaries (totals, means) but cannot identify neighbor-based purchase patterns for personalized suggestions.

6) Conclusion
Collaborative filtering via data mining is the appropriate approach; this confirms the correct answer: Use data mining with collaborative filtering to identify patterns across similar customers.

类似问题

Question19 In Collaborative Filtering, what is considered among the drawbacks of this type of recommender system Not accurate in comparison to other recommender systems Can not predict properly for people with unique tastes Slow at query time ResetMaximum marks: 0.5 Flag question undefined

When recommending a show to a user, collaborative filtering will pay more attention to:

What does collaborative filtering automate in large datasets?

Collaborative filtering is can be used to _____.

Collaborative filtering recommends items based on patterns from “users similar to you”, which makes it more advanced at preventing algorithmic bias such as review polarization.

Collaborative Blank ______ is a technique used in recommendation systems to provide personalized recommendations to users based on their similarity to other users.

What is a technique used in recommendation systems to provide personalized recommendations to users based on their similarity to other users?

The implementation and loading phase of the Database Life Cycle (DBLC) involves _____.

更多留学生实用工具

智能学习助手

风格化写作助手

论文查重助手

文献引用助手

课堂转译助手

课堂笔记助手

Quiz搜索助手

学校历年真题

智能刷题助手

智能匹配练习

希望你的学习变得更简单