IPA IP 2023 Question 6
A question IPA Japan published in 令和5年度 ITパスポート試験 公開問題 問6, translated into English. ITPEC is based on the same exams.
Company A acquires data consisting of many items, such as customer behavior, the weather, and the locations of stores. It is considering explaining changes in sales quantity by analyzing these data. In doing so, it wants to narrow down the parameters used for the explanation to as few as possible. Which of the following is the most appropriate analysis method to be used in this case?
Explanation
Principal component analysis — it combines many correlated items into a few components that explain most of the variation.
Principal component analysis takes data with many variables (customer behavior, weather, store location and so on) and builds new composite variables, the principal components, each a weighted combination of the originals. The first component captures as much of the data's variation as possible, the second as much of the rest, and so on. Keeping only the first few components explains the change in sales with a small number of parameters, which is what Company A wants: dimensionality reduction.
Answer (c)
Why not others:
- •(a) ABC analysis — ranks items by their share of sales or value (a Pareto chart) to decide which to manage closely; it does not reduce many explanatory items to a few
- •(b) Cluster analysis — groups similar customers or products together by their attributes; it classifies records rather than summarizing variables
- •(d) Correlation analysis — measures the strength of the relationship between two variables at a time; it does not merge many items into fewer parameters
Key rule: To summarize many variables with a few composite indicators, use principal component analysis; to group similar records, cluster analysis; to measure how two variables move together, correlation analysis.
Corporate Activities · AI-written, may contain errors
This question comes from the past questions IPA Japan publishes, translated into English by ITPEC Practice. ITPEC Practice is an independent study tool and is not affiliated with IPA or ITPEC. See the original paper (PDF) or Report an issue.