cxxr / livestats Goto Github PK

Apache Commons Math Stats supports setting a window size for rolling stats. Is there a way to set the window size in LiveStats?

The bot created this issue to inform you that pyup.io has been set up on this repo.
Once you have closed it, the bot will open pull requests for updates as soon as they are available.

[1, 5, 6, 7, 9, 12, 15, 19, 20]
for this set LiveStats shows me 8.68209876543 but this is not a median for this values! It's not just not-accurate - this just not a median. Just becouse median must split set to 2 parts with same length, this value wrong.
Any ideas why this happened and how to solve this?

Try out this simple code which gives 10 values in [60-61) and then keeps giving values only between [50-51)

After 100 total samples, we expect theoretically that the percentile value will be around in (50-51). The As P^2 is just an estimate, I break out of the loop once we get a value of less than 52. Now this takes anywhere from 5000 samples forever.

Try running this simple code multiple times and see..

from livestats import livestats
from math import sqrt
import random

low = 50
high = 50*1.2
randomdata = iter(random.random, 1)

test = livestats.LiveStats([0.99])
count = 0;
for count in xrange(10):
        test.add(randomdata.next() + high)

for count in xrange(50000000):
        test.add(low)
        if(count%100 == 0): 
                print "count {}: Average {}, stddev {}, quantiles {}".format(count, test.mean(), sqrt(test.variance()), test.quantiles())
        if(test.quantiles()[0][1] < low+2):
                break;

print "Done: count {}: Average {}, stddev {}, quantiles {}".format(count, test.mean(), sqrt(test.variance()), test.quantiles())