Hi, Could someone please guide me as to what the threshold factor in the filloutliers represent?

I need to understand if I have a dataset (Data) of (N,M) rows and columns and I use filloutliers to identify the outliers then what does the threshold factor represent in the following script:

filloutliers(Data,'center','median','ThresholdFactor',1.5)

According to mathworks guide the threshold factor is defined as the scaled median absolute deviation, but I need some theoretical background as to what setting 1.5 threshold factor mean? Thanks!

Answer 1

[Tridib]

Hi @Akbar Raza,

Outliers are data points that differ significantly from most others in a dataset, such as a 50-year-old in a class where most students are 20–22. The Median Absolute Deviation (MAD) method identifies outliers by measuring how far each value is from the median:

• Find the center (median) of the data.
• Measure how far each point is from the center (absolute deviations).
• Find the median of these deviations (MAD).

To decide what counts as an outlier, a threshold factor (like 1.5) multiplies the MAD, and any value further than this scaled distance from the median is flagged as an outlier. This threshold lets you control how strict the definition is, helping to distinguish between normal variation and truly unusual values by adjusting the "normal range" around the median.

For more help, refer to the following documentation:

https://www.mathworks.com/help/matlab/ref/filloutliers.html

Hope this helps!