search

pennekampster / bemovi

Bemovi, software for extracting BEhaviour and MOrphology from VIdeos
5 stars 11 forks source link

trajectory summary #5

Closed pennekampster closed 9 years ago

pennekampster commented 9 years ago

We could use the median instead of the mean; probably good idea to check some of the within trajectory variation

opetchey commented 9 years ago

I don't understand this. Can you expand?

pennekampster commented 9 years ago

It's just that the mean would be more strongly affected by, say an erroneous measurement of size due to merging of an individual with a large debris particle, than the median. I am also not sure all traits we average are normally distributed.

opetchey commented 9 years ago

I see. Agree. Suggest we put a switch in the function, for summarise with median or mean, with median as the default.

pennekampster commented 9 years ago

Issue fixed by Marco.