Outlier detection window

  1. Outlier detection window
    1. Selecting outliers with the mouse
    2. Zooming in and out
    3. Detecting outliers automatically

The outlier detection window allows you to mark outliers and remove them from further analysis. As shown below, outlier detection operates on a plot of the data, one plot for each type of measurement (absorbance, luminescence, and fluorescence). Valid data points are blue, whereas outliers are surrounded by a red circle.

If there are several measurements of each type, then a listbox allows you to switch between those.

Selecting outliers with the mouse

The first way to select outliers is manually, by means of the mouse. This can be done in several ways:

  1. To select or deselect a data point, left-click on it.
  2. To select a group of points, draw a rectangle around them with the mouse, while pressing the left-button (the cursor becomes a small cross with a + sign)
  3. To deselect a group of points, do the same as to select them, but while pressing the right button (or the ALT key). The cursor is a small cross with a - sign.

Zooming in and out

Sometimes data points are too close to be clearly distinguishable. In these cases, you can zoom in or out in one of the following ways:

  1. By right-clicking a menu appears that allows you to zoom by a factor of 2 or 0.5, or to reset the view.
  2. You can drag a rectangle with the mouse, while holding the SHIFT key. The cursor changes to a magnifying glass with a cursor inside.

Detecting outliers automatically

If there are many outliers, manual detection may become cumbersome. In this case, you can use the automatic outlier detection procedure, based on iterative fits of a regression spline. After each fit, the data with the highest residuals are removed. The user can specify some parameters to control this procedure, notably the order of the spline, the number of knots, and a lower limit on the percentage of remaining points.