TMDDataFormat Type

Summary

Defines MD matrix data type.

Declaration

TMDDataFormat = (mdFormatDissimilarities,mdFormatSimilarities,mdFormatRaw);

Value	Description
mdFormatDissimilarities	Dissimilarities represent the distance between two objects. They may be measured directly or approximated. MDS algorithms use the dissimilarities directly. A dissimilarity matrix is always symmetrical and with zero main diagonal.
mdFormatSimilarities	Similarities represent how close (in some sense) two objects are. Similarities must obey the rule: similarity(i,j) <= similarity(i,i) and similarity(j,j) for all i and j. Similarity matrices are symmetrical. Similarities are converted to dissimilarities by the following relation: d(i,j) = Sqrt[s(i,i) + s(j,j) -2*s(i,j)] where d(i,j) represents a dissimilarity and s(i,j) represents a similarity. In case your data consists of standard measures rather than dissimilarities or similarities, you can create a similarity matrix by creating the correlation matrix.

Value

Description

mdFormatDissimilarities

Dissimilarities represent the distance between two objects. They may be measured directly or approximated. MDS algorithms use the dissimilarities directly. A dissimilarity matrix is always symmetrical and with zero main diagonal.

mdFormatSimilarities

Similarities represent how close (in some sense) two objects are. Similarities must obey the rule:

    similarity(i,j) <= similarity(i,i) and similarity(j,j)

for all i and j. Similarity matrices are symmetrical. Similarities are converted to dissimilarities by the following relation:

    d(i,j) = Sqrt[s(i,i) + s(j,j) -2*s(i,j)]

where d(i,j) represents a dissimilarity and s(i,j) represents a similarity. In case your data consists of standard measures rather than dissimilarities or similarities, you can create a similarity matrix by creating the correlation matrix.