Manrotate: Difference between revisions
imported>Jeremy (Importing text file) |
imported>Jeremy (Importing text file) |
||
| Line 1: | Line 1: | ||
===Purpose=== | ===Purpose=== | ||
Graphical interface to manually rotate model loadings and investigate directions in the scores. | Graphical interface to manually rotate model loadings and investigate directions in the scores. | ||
===Synopsis=== | ===Synopsis=== | ||
:manrotate(model,''lvs'') | :manrotate(model,''lvs'') | ||
===Description=== | ===Description=== | ||
MANROTATE shows a score vs. score scatter plot and model loadings and allows the user to "rotate" the loadings. The loadings (shown as two colored lines in the score/score plot) can be dragged through different angles observing the resulting loading shape in the loadings plot (Loadings are always kept orthogonal.) | MANROTATE shows a score vs. score scatter plot and model loadings and allows the user to "rotate" the loadings. The loadings (shown as two colored lines in the score/score plot) can be dragged through different angles observing the resulting loading shape in the loadings plot (Loadings are always kept orthogonal.) | ||
This interface is useful to identify a loading "shapes" which point towards, and orthogonal to, a given sample cluster or direction. | This interface is useful to identify a loading "shapes" which point towards, and orthogonal to, a given sample cluster or direction. | ||
The user clicks on the heavy lines in the scores plot and "drags" them to point in a selected direction. The loadings (shown on the right in the figure) are automatically updated to show the loading which accounts for the new direction in the scores plot. The rotated loading vectors can be saved to the workspace using the toolbar save button. | The user clicks on the heavy lines in the scores plot and "drags" them to point in a selected direction. The loadings (shown on the right in the figure) are automatically updated to show the loading which accounts for the new direction in the scores plot. The rotated loading vectors can be saved to the workspace using the toolbar save button. | ||
Inputs include a PCA, PLS, PCR, or other 2-way factor-based model, model, and an optional input, ''lvs'', which is a two-element vector specifying which of the model factors should be plotted and rotated (default = [1 2] which plots factor 2 vs factor 1.) | Inputs include a PCA, PLS, PCR, or other 2-way factor-based model, model, and an optional input, ''lvs'', which is a two-element vector specifying which of the model factors should be plotted and rotated (default = [1 2] which plots factor 2 vs factor 1.) | ||
===See Also=== | ===See Also=== | ||
[[pca]], [[pcr]], [[pls]], [[varimax]] | [[pca]], [[pcr]], [[pls]], [[varimax]] | ||
Latest revision as of 15:25, 3 September 2008
Purpose
Graphical interface to manually rotate model loadings and investigate directions in the scores.
Synopsis
- manrotate(model,lvs)
Description
MANROTATE shows a score vs. score scatter plot and model loadings and allows the user to "rotate" the loadings. The loadings (shown as two colored lines in the score/score plot) can be dragged through different angles observing the resulting loading shape in the loadings plot (Loadings are always kept orthogonal.)
This interface is useful to identify a loading "shapes" which point towards, and orthogonal to, a given sample cluster or direction.
The user clicks on the heavy lines in the scores plot and "drags" them to point in a selected direction. The loadings (shown on the right in the figure) are automatically updated to show the loading which accounts for the new direction in the scores plot. The rotated loading vectors can be saved to the workspace using the toolbar save button.
Inputs include a PCA, PLS, PCR, or other 2-way factor-based model, model, and an optional input, lvs, which is a two-element vector specifying which of the model factors should be plotted and rotated (default = [1 2] which plots factor 2 vs factor 1.)