![[attach.png]](idldoc-resources/attach.png)
./
detectoutlier.pro
Routines
detectoutlier_plot, x, y, resultPlot which objects were flagged as outliers from detectoutlier.
result = detectoutlier(x, y, status [, THRESH=THRESH] [, VERBOSE=VERBOSE] [, PLOT=PLOT])This function attempts to identify outlier data points in 2D space.
Routine details
topdetectoutlier_plot
Display utilities
detectoutlier_plot, x, y, result
Plot which objects were flagged as outliers from detectoutlier.
Parameters
- x in required
Input x points
- y in required
Input y points
- result in required
A vector of [0,1] to denote [outlier/non-outlier]
topdetectoutlier
statistics
result = detectoutlier(x, y, status [, THRESH=THRESH] [, VERBOSE=VERBOSE] [, PLOT=PLOT])
This function attempts to identify outlier data points in 2D space. It proceeds as follows: 1) Calculate the center position of all points 2) Remove the biggest outlier, and recalculate the center and standard deviation. 3) If the outlier removed in step 2 deviates from the new center by more than THRESH * stdev, go to step 1 and repeat. Otherwise, re-insert the outlier flagged in step 2 as a legitimate data point, and return.
Parameters
- x
- y
- status
Keywords
- THRESH in optional
Use to manually set the threshhold for outlier detection. The default is 3.5
- VERBOSE in optional
Output textual information
- PLOT in optional
Plot the points, color coded by whether they are outliers or not.
File attributes
| Modifcation date: | Mon Mar 22 16:17:13 2010 |
| Lines: | 142 |