2026-03-06T02:38:08Z https://kansai-u.repo.nii.ac.jp/oai

oai:kansai-u.repo.nii.ac.jp:00018527 2023-05-15T20:46:53Z 528:1385:1386:1390

非対称データに基づく地域構造分析のアプローチ : Structural Modelingによる多次元尺度構成法 ONE APROACH TO REGIONALIZATION FOR ASYMMETRIC SCALING DATA : Multi-dimensional scaling by Structural Modeling 辻, 光宏 多次元尺度構成法 非対称 ISM 地域間分析 電話発信量 multi-dimensional scaling asymmetric ISM regionalization number of telephone calls 関西大学 Kansai University 多次元尺度構成法によって地域間分析をおこなう時に，地域間の類似度が一意に決定できない場合が多く，分析が複雑になる。例えば，我が国の47都道府県を地域と考え，NTT電話回線の発信量（総通信時間）を元にして地域間分析する場合には，地域iから地域jへの電話発信量に基づく依存度を地域iから地域jへの類似度Sijと仮定すると, ､SijとSjiとが等しくなく，類似関係を示す正方行列が非対称となる。この報告では，社会システム理論の一手法であるISM (Interpretive Structural Modeling)を用いて，電話による情報流通の地域構造を相互依存の観点から同定し，さらに，この同定結果を用いて多次元尺度構成法による幾何学的な地域構造表現をアプローチする。 Multi-dimensional scaling can de employed to represent each stimulus between different regions as a point in a small dimensional space. Most regional analysis are not capable of identifying similarities between regions. As one example of the kind of situation in which multi-dimensional scaling is appropriate, and assuming that each of the 47 provinces of Japan is a separate region, in the analysis of the number of telephone calls between these different areas, the differences between the responsibility of Region A on Region B and the responsitility of Region B on Region A can be easily determined with multi-dimensional scaling. Several methods have been proposed to analyze asymmetric proximities, and two steps methods are presented here. The first step is a process involving the application of Interpretive Structural Modeling (ISM) graph theory, in which we analyze grouping status (as delimited regions). The next process is Multi-dimensional Scaling. 関西大学社会学部 1992-09-30 jpn departmental bulletin paper VoR http://hdl.handle.net/10112/00022575 https://kansai-u.repo.nii.ac.jp/records/18527 AN00046982 02876817 関西大学社会学部紀要 24 1 71 88 https://kansai-u.repo.nii.ac.jp/record/18527/files/KU-1100-19920930-03.pdf application/pdf 2.1 MB 2021-02-01