報告書番号

R03025

タイトル（和文）

効果的なアンケート設計手法 -高精度なデータ分析のための調査項目選択-

タイトル（英文）

Effective questionnaire design method - selection of effective questionnaire items

概要 （図表や脚注は「報告書全文」に掲載しております）

アンケート調査を実施する場合, 調査項目数が増えると回収率の低下や調査コストの増大となる. そこで, 少ない調査項目数で, より高い精度の知識が抽出できる効率的なアンケートを設計する手法を開発し、効果を検証した。

概要 （英文）

Data-mining technology to extract knowledge from large database within a corporation is an essential tool for today's enterprise. It is essential to prepare effective database before data-mining, because we cannot extract effective knowledge without it. The questionnaire is one of common ways to collect data. However, the size of feasible questionnaires is limited by budget, because the questionnaire costs more as the number of questions increases. We have to design an effective questionnaire for accuracy improvement of the extracted knowledge within budget. In this paper, we extended `critical data specification method' in order to design an effective questionnaire with a predefined number of questions. We applied the proposed method to the problem of constructing a database for prediction of `Minryoku level' of municipalities. The proposed method can identify a set of attributes that effectively improves accuracy of the extracted knowledge by addition of these attributes.

報告者

キーワード