Closed amritbhanu closed 5 years ago
here, clearly 0.2 is too big but after that, no evidence of improvement below 0.1 which means for n= log(1-c)/log(1-(e^2)), our e = 0.1
please draw using my xtile format. if drawn like the following, then your display for a research paper would be very small. also, you'd get stats results for each epsilon
what happens with executions over time? when are their demonstrable improvements at epsilon= 0.1?
-
you using my stats.py for the one? what is the Cohen value here? I ask since some of your "differences" are crazy small. eng in poi medians of 742 are being scored ranks 2 AND 3
t
using your stats.py, cohen=0.3 median is 742 but check the iqr for that.
here to, set epsilon to 2
camel
Jedit
Poi
log4j
synapse
velocity
xalan
xerces
ivy
lucene