One of the types of objects in PRAAT.
An object of type Configuration represents the positions of a number of labelled points in a multidimensional space.
How to create a Configuration
From the New menu:
-
• Create Configuration...
By multidimensional scaling:
-
• Dissimilarity: To Configuration (monotone mds)...
-
• Dissimilarity: To Configuration (i-spline mds)...
-
• Dissimilarity: To Configuration (interval mds)...
-
• Dissimilarity: To Configuration (ratio mds)...
-
• Dissimilarity: To Configuration (absolute mds)...
By multidimensional scaling with weights (Dissimilarity & Weight: To Configuration...):
-
• Dissimilarity & Weight: To Configuration (monotone mds)...
-
• Dissimilarity & Weight: To Configuration (i-spline mds)...
-
• Dissimilarity & Weight: To Configuration (interval mds)...
-
• Dissimilarity & Weight: To Configuration (ratio mds)...
-
• Dissimilarity & Weight: To Configuration (absolute mds)...
By multidimensional scaling with a start Configuration:
-
• Dissimilarity & Configuration: To Configuration (monotone mds)...
-
• Dissimilarity & Configuration: To Configuration (i-spline mds)...
-
• Dissimilarity & Configuration: To Configuration (interval mds)...
-
• Dissimilarity & Configuration: To Configuration (ratio mds)...
-
• Dissimilarity & Configuration: To Configuration (absolute mds)...
By transforming an existing Configuration:
-
• Configuration: To Configuration (varimax)...
-
• Configuration & AffineTransform: To Configuration
-
• Configuration & Procrustes: To Configuration
From Principal component analysis:
-
• TableOfReal: To Configuration (pca)...
-
• PCA & TableOfReal: To Configuration...
From Discriminant analysis:
-
TableOfReal: To Configuration (lda)...
-
• Discriminant & TableOfReal: To Configuration...
How to draw a Configuration
How to modify a Configuration
Inside a Configuration
With Inspect you will see the following attributes:
-
numberOfRows
-
the number of points (numberOfPoints≥1).
-
numberOfColumns
-
the dimension of the space (numberOfDimensions≥1).
-
rowLabels
-
the names associated with the points.
-
columnLabels
-
the names for the dimensions.
-
data [1..numberOfPoints] [1..numberOfDimensions]
-
the coordinates of the points.
-
metric
-
determines the way distances between points are measured. In general the distance between points xi and xj is:
dij = (∑k=1..numberOfDimensions wk |xik – xjk|metric)1/metric |
-
For Euclidean distances metric is 2.
-
w [1..numberOfDimensions]
-
weight given to each dimension in the distance calculation.
Links to this page
© djmw, November 2, 2010