IPA IP 2020 Question 4
A question IPA Japan published in 令和2年度 10月 ITパスポート試験 公開問題 問4, translated into English. ITPEC is based on the same exams.
Company A, which runs a nationwide chain of convenience stores, has accumulated at its head office detailed sales data of each store over the past 10 years. The company wants to build a system that analyzes the relationship between this sales data, weather data for the past 10 years, and information on events in the neighborhood of each store, grasps the correlation between weather conditions, event information, and the sales volume of products, and forecasts the sales of each store with higher accuracy from weather forecasts up to one week ahead and event information. Which of the following is the most appropriate technology to be used in this case?
Explanation
AI technology such as deep learning — learning from years of data how weather and events relate to sales, and using that to forecast sales.
The company already holds the data: ten years of detailed sales for every store, ten years of weather data and local event information. What it needs is a way to find the relationships hidden in that large, many-factored data and turn them into forecasts. Machine learning, deep learning in particular, does exactly that: a model is trained on the past data to learn how weather conditions and events affect the sales of each product, and is then given the coming week's weather forecast and events to predict each store's sales. The more data it learns from, the more accurate the forecasts tend to become.
Answer (c)
Why not others:
- •(a) — IoT collects data automatically from sensors and devices, but here the data is already accumulated; the task is analyzing it and forecasting, not collecting it
- •(b) — virtual reality lets people experience a virtual space as if it were real; it has nothing to do with analysis or forecasting
- •(d) — spreadsheet statistics suit small, simple analyses; finding the patterns in ten years of detailed sales for every store across many factors, and forecasting more accurately, is a task for AI
Key rule: When large accumulated data must be analyzed to learn patterns and predict, the answer is AI (machine learning, deep learning); IoT gathers data, it doesn't predict from it.
Business Industry · 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.