The analysis of proximities: Multidimensional scaling with an unknown distance function. I.
Citations
12,519 citations
6,875 citations
5,002 citations
Cites methods from "The analysis of proximities: Multid..."
...The similarity ratings were scaled by M-D scale (Shepard, 1962; Shepard, Romney, & Nerlove, Vol. 1, 1972)....
[...]
4,741 citations
4,561 citations
Cites methods from "The analysis of proximities: Multid..."
...In a companion paper [7] we describe the rationale for our approach to scaling, which is related to that of Shepard [ 9 ]....
[...]
References
5,419 citations
2,920 citations
"The analysis of proximities: Multid..." refers background in this paper
...The fact that correlations or covariances can be interpreted as scalar products of vectors ([ 31 ], pp. 89-91) turns out to provide a much stronger starting position than is possible in the analysis of proximities with an unknown distance function....
[...]
1,984 citations
"The analysis of proximities: Multid..." refers methods in this paper
...There are many examples in this class: Richardson's method of triadic combinations [24], Klingberg's multidimensional method of rank orders [19], Torgerson's complete method of triads ([ 32 ]; [33], pp. 263-268), and Messick and Abelson's method of successive intervals [1, 22]....
[...]
1,005 citations
"The analysis of proximities: Multid..." refers background or methods in this paper
...In fact the computer program described here can be regarded, in the terminology of Stevens ([ 30 ], p. 25), as an automatic method for essentially transforming the given "ordinal scale" of similarities into a "ratio scale" of distances....
[...]
...a unit of measurement [5], it has been widely recognized that much can be accomplished even when the measuring operations are not sufficiently quantitative to yield what Stevens [ 30 ] has termed an "interval scale." Following Coombs, scales obtained by these less quantitative operations have been called "ordered metric scales" because the operations usuutly determined partial ordering on the distances between stimuli....
[...]
...In such cases an ordered metric scale is stronger than an ordinal but weaker than an interval scale (el., [ 30 ], p. 25)....
[...]
...In this case the analysis begins, essentially, with a rank ordering of 105 proximity measures and yields, as output, 30 coordinates on axes with the properties of an interval scale (see [ 30 ], p. 25)....
[...]