How to remove bad data (outliers) from a plot?

I plot my data and a line of best fit together. I realized there are a few obvious bad data (outliers) in my plot and I need to remove them.

How do I do that?

Thanks!

[STEPHEN ROSS]

For small data sets less than 60, use Peirce's Criterion for an excellent statistical method: "https://www.researchgate.net/publication/267241891_Peirce%27s_criterion_for_the_elimination_of_suspect_experimental_data"

Answer 1

[Image Analyst]

See Brett's File Exchange submission: http://www.mathworks.com/matlabcentral/fileexchange/3961

Answer 2

[Star Strider]

The most acceptable way is to set confidence limits based on a particular probability and eliminate those data that are outside them. (If you have a large data set, you would expect some valid data to be outside the expected confidence limits.) Various Statistics Toolbox functions such as regress for linear regression and nlpredci for nonlinear regression can calculate these confidence limits for you.

Answer 3

[Steven Lord]

If you're using release R2017a or later, use the isoutlier and/or filloutliers functions in the "Missing Data and Outliers" section on this documentation page.