A Nonparametric Outlier Rejection Scheme
Author Information
Author(s): j. s. o. Odonde
Primary Institution: P.O. Box 151, 4530 AD Terneuzen, The Netherlands
Hypothesis
The paper proposes a new method for removing outliers from small datasets using a nonparametric approach.
Conclusion
The study presents an effective outlier rejection routine based on bounding data points to within the mean absolute deviation from the mode.
Supporting Evidence
- The outlier rejection algorithm is based on bounding data points to within the mean absolute deviation from the mode.
- The study presents two sets of experimental data to illustrate the performance of the outlier rejection algorithm.
- The mean absolute deviation is recognized as a more robust estimate of the width around the central tendency.
Takeaway
This study shows a way to clean up messy data by finding and removing the bad points that don't fit with the rest.
Methodology
The method involves calculating the mode, bounding data points, detrending data, applying a Fast Fourier Transform, low pass filtering, and restoring the mean.
Potential Biases
The first and last data points may come from previous or subsequent experiments, introducing bias.
Limitations
The method may reintroduce unwanted characteristics from previous experiments when restoring the linear trend.
Want to read the original?
Access the complete publication on the publisher's website