問4
頻出度★★☆
コンビニエンスストアを全国にチェーン展開するA社では, 過去10年間にわたる各店舗の詳細な販売データが本部に蓄積されている。これらの販売データと, 過去10年間の気象データ, 及び各店舗近隣のイベント情報との関係を分析して, 気象条件, イベント情報と商品の販売量との関連性を把握し, 1週間先までの天気予報とイベント情報から店舗ごとの販売予想をより高い精度で行うシステムを構築したい。このとき活用する技術として, 最も適切なものはどれか。
令和2年度 ITパスポート試験 問4
テクノロジ系基礎理論 › 情報に関する理論
This question's explanation and your answers so far.
解説
問 4
Incorrectア
Technology for automatically collecting data from sensors and the like using IoT is a technology for gathering real-world data. The purpose of this question is to analyze and predict from a large amount of already-accumulated data, and collecting new data is not the challenge, so it is not the most appropriate.
Incorrectイ
Virtual reality (VR) technology provides real-like experiences in a virtual space and is used for games, training simulations, and the like. It is unrelated to data analysis or sales prediction and does not match the purpose of this question.
- ウ
正解 ウ is correct. The essence of this case is to combine the vast 10 years of sales data already accumulated at headquarters with weather data and event information, analyze the complex correlations between them and sales volume, and predict sales volume with high accuracy from the weather forecast and event information. AI technology such as deep learning (machine learning) can learn features, patterns, and correlations on its own from large amounts of data without a human explicitly providing rules, and make highly accurate predictions for unknown inputs. Therefore, the most suitable technology for this system, whose purpose is to grasp the relationships in large data and predict the future, is ウ.
Incorrectエ
Statistical analysis with spreadsheet software can be used for simple aggregation and basic statistical processing, but it is not powerful enough to grasp complex multivariate correlations and make high-accuracy predictions over large-scale data spanning many stores nationwide and 10 years, so it cannot be called the most appropriate.
ここがポイント
- For applications that learn correlations and patterns from large amounts of accumulated data to predict the future, AI technology such as deep learning (machine learning) is suitable.
- Choose technology by purpose: IoT = data collection, VR = virtual experience, spreadsheets = simple aggregation and statistics, AI / deep learning = analysis and prediction of large-scale data.
用語