Goodreads helps you follow your favorite authors. Be the first to learn about new releases!
Start by following Agustin Lebron.
Showing 1-1 of 1
“A surprisingly good heuristic for "how outlier-y/crappy is my data":
- N is number of datapoints.
- Sort your data.
- Get rid of the first and last k * sqrt(N) points.
For what value of k does the data start to look "sensible"?
k < 0.3: pretty good data
k > 3: pretty bad data”
―
- N is number of datapoints.
- Sort your data.
- Get rid of the first and last k * sqrt(N) points.
For what value of k does the data start to look "sensible"?
k < 0.3: pretty good data
k > 3: pretty bad data”
―
