CODEWITHSUNDEEP
data-miningwekaoutliers
theQuestion
theQuestion

Reputation: 3

How can Weka be used to find outlier values based on SD?

I want to use Weka to find outliers by locating the attribute values more than 3 standard deviations from the mean. I'm sure there's a way to do it.

I already know how to search for outliers using the InterquartileRange filter--that just isn't what I want to do in this case. And Weka gives me the the mean and SD, so figuring out my thresholds is easy enough. I just don't know what filter to use, and searching hasn't worked (I get answers to much more advanced questions).

Upvotes: 0

Views: 249

Answers (1)

fracpete
fracpete

Reputation: 2608

If you compute the thresholds yourself, you could use the SubsetByExpression filter to discard instances that fall outside these values (for a particular attribute).

Upvotes: 0

Related Questions