Table of Contents
Metric Compute Confidence Ellipse
Overview
The Metric Compute Confidence Ellipse command can be used to compute the 95 % confidence ellipse in 2 or more dimensions. This is the smallest ellipse that will cover 95 % of the data points.
Can also be used to compute the area or volume of an ellipse based on PCA. Does not require all vertices to be present for all frames. It can be found within the Pipeline Workshop under Metric.
Pipeline Command
The syntax for the function is as follows:
Metric_Compute_Confidence_Ellipse /SIGNAL_TYPES= ! /SIGNAL_FOLDER=ORIGINAL ! /SIGNAL_NAMES= ! /COMPONENT_SEQUENCE= ! /RESULT_FOLDER=PROCESSED /RESULT_NAME= ! /EVENT_SEQUENCE= ! /EXCLUDE_EVENTS= ! /START_FRAME= ! /END_FRAME= ;
Command Parameters
The Parameters for the function are as follows:
| /SIGNAL_TYPES= | Specify signal type |
| /SIGNAL_NAMES= | Specify signal name |
| /SIGNAL_FOLDER= | Specify signal folder |
| /COMPONENT_SEQUENCE | Which component of the signal will be used for statistical analysis of the sequence. |
| /RESULT_FOLDER | Name of folder holding resulting processed data |
| /EVENT_SEQUENCE= | Specify the sequence of Events. Any number of Events can be entered (separated by +). This specific sequence of events must be true for a metric to be computed. The metric is computed from the first event in the sequence to the last event in the sequence |
| /EXCLUDE_EVENTS= | If this event occurs before the first and last event, do not compute a metric |
| /START_FRAME= | The frame at which to start recording data. |
| END_FRAME= | The frame at which to stop recording data. |
Dialog
The command can be edited in a text editor or in a dialog form. To edit in the dialog pop up form either click on the Edit button in the pipeline workshop or double-click on the pipeline command. The dialog is shown below.
The dialog box allows you to assign values to the command parameters outlined above.
Example: Computing Confidence Ellipse for COFP
In this example the Metric Compute Confidence Ellipse command is used to compute the values associated with the confidence ellipse for left and right COFP during a balance trial. This analysis can be useful to compare balance consistency between sides using ellipse areas.
The pipeline command for each side is:
Metric_Compute_Confidence_Ellipse /SIGNAL_TYPES=COFP ! /SIGNAL_FOLDER=ORIGINAL /SIGNAL_NAMES=FP3 /COMPONENT_SEQUENCE=X + Y ! /RESULT_FOLDER=PROCESSED /RESULT_NAME=Ellipse_Right /EVENT_SEQUENCE=START+END ! /EXCLUDE_EVENTS= ! /START_FRAME= ! /END_FRAME= ; Metric_Compute_Confidence_Ellipse /SIGNAL_TYPES=COFP ! /SIGNAL_FOLDER=ORIGINAL /SIGNAL_NAMES=FP4 /COMPONENT_SEQUENCE=X + Y ! /RESULT_FOLDER=PROCESSED /RESULT_NAME=Ellipse_Left /EVENT_SEQUENCE=START+END ! /EXCLUDE_EVENTS= ! /START_FRAME= ! /END_FRAME= ;
The resulting metric is found in the METRIC:PROCESSED folder and should look similar to below.
The metric for this example contains 3 components, with 3 values for each component. The results are as follows:
Frame 1:
| X | Volume |
| Y | average of all frames of input component 1 |
| Z | average of all frames of input component 2 |
Frame 2: (for input component 1)
| 1 | Eigenvalue |
| 2 | Eigenvector |
| 3 | Eigenvector |
Frame 3: same as frame 2, but for input component 2
If you have two input components, frame 1 X is the area, if you have three input components, Frame 1, X is the volume.
The ellipse can also be visualized in the reports tab:
Add a new 2D Graph to your report with the appropriate X and Y data and make sure to select Graph Confidence Ellipse.
The resulting graph should look similar to below:
Notes
For more information see, confidence hyper ellipse eigenvalues.