0

I have an array of measurements, some of those measurements are faulty, so if I do a min-max search, I get those faulty values.

I'm trying to filter them out by finding 1% min and 99% max quantiles.

I don't want to sort the data

How can I find the value at which 99% of the data is above and value at which 99% of the data is below?

Can I do it using weighted averages?

J. W. Tanner
  • 60,406
Mich
  • 201
  • If you have $N$ measurements, then you want to find the value at which $N_0=0.01*N$ of the measurements are below/above that value. So you could just compute $N_0$ and go through your list of measurements (without sorting) and filter out the $N_0$ smallest/largest values and determine possible values from this filtered out list? – ym94 Dec 27 '23 at 21:00
  • 1
    You can use a linear-time selection algorithm. – Karl Dec 27 '23 at 21:54

0 Answers0