問5
頻出度★★★
経営判断や業務遂行などに活用するために, データベースに蓄積された顧客の購買履歴などのデータを分析して, 有用な情報を抽出する手法はどれか。
平成24年度 秋期 ITパスポート試験 問5
ストラテジ系企業活動 › 業務分析・データ利活用
This question's explanation and your answers so far.
解説
問 5
Incorrectア
Data cleansing is the preprocessing work of detecting and correcting inconsistent notation, duplicates, missing values, and miswritten entries in data, so that the data reaches a quality that can stand up to analysis. It is not a method for analyzing data to extract useful information, but a step that cleans the data up before the analysis, so it is wrong.
- イ
正解 The correct answer is “イ Data mining.” Just as the word mining means “digging out,” data mining is a method that analyzes the large volumes of data accumulated in a database or a data warehouse using techniques such as statistical analysis and machine learning, and digs out the useful knowledge buried there: regularities, correlations, and trends that had not been noticed before. The question's phrases “analyze data such as customers' purchase histories and extract useful information” and “use it for management decisions, business operations, and so on” describe exactly this purpose and procedure. A typical example is finding an association rule such as “customers who buy product A also tend to buy product B” (basket analysis), which is then put to use in shelf layout, recommendations, and sales promotion measures. Note also that a database dedicated to analysis, in which the data to be analyzed is organized and integrated in time series and accumulated, is called a data warehouse; data mining is the analysis method carried out on that warehouse.
Incorrectウ
Data managing is a way of referring to the whole activity of managing data resources at the organizational level: collecting, storing, and protecting data, and establishing rules for putting it to use. It is a management framework, not a term for an analysis method that extracts regularities or useful information from accumulated data, so it is wrong.
Incorrectエ
Data modeling is the work of organizing the structure of the data handled in the business and the relationships among it, using E-R diagrams and the like, to produce a blueprint for the database (a data model). It is the design done before the database is built, not a method for analyzing already accumulated data to take information out of it, so it is wrong.
ここがポイント
- Data mining = a method that analyzes large volumes of accumulated data with statistical analysis and machine learning, discovers the regularities and correlations buried in it (for example, a tendency to buy two items together), and puts them to use in management decisions. Learn it as a pair with the data warehouse, where the data for analysis is accumulated.
- Distinguish the roles of the confusingly similar “data …” terms: data modeling = design (before the analysis); data cleansing = preprocessing that gets the quality in order; data mining = the stage of analyzing and extracting knowledge.
用語