ok, hi

hi

greeting to Poland

a am Slovak

thx, same to you :)

well, i want to avoid spacing my data to hours or days

so, generally, what you want to ask?

why?

i will lose much information by this process

for example, the activity of bats is like wave like pattern

it looks like a sinusoid

series of 0,0,1,1,0... is also a series, so you don't have to - it depends on the method you want to use

ok, but consider a situation

if you want to threat it as binomial, then you have n,k parameters so a NUMBER of successes in N trials

so you need some frame for it

aha

also 0,0,1,1,0 is equivalent to 2 successes in 5 trials

aha forget binomial

unless you are interested in the position in sequence - but this depends on method you want to use

may I send you my article?

i will enlighten the whole situation

feel free, t.wolodzko@ibe.edu.pl

thank you

ok, I got it

please look at figures 1 and 2

the general Q is: what do you want to do with such series..?

this behaviour (wave like pattern of alternating periods of high and low activity) is present on multiple locations is space

it is a signalling behaviour

i have recording from multiple locations during one night

only one location will be used by bats to use it a a roost

sequences are very similar among locations

ok, I see it, but what do you want to do with it?

question: in what parameter is sequence of activity propagated at locality chosen by bats different to others ?

btw way i am zoologist, not a trained statistician

so do not laugh :D

np, me neither

you probably need some more precise operationalization in here

it is not a problem to use a series of 0 and 1 as a time series

ok

it is also not a problem to sum them (most time-series are some kind of aggregates)

the question is - what do you want to achieve?

it sounds rather as if you were not sure how to analyze such data

try googling "binary time-series" - it provides multiple references

I want to be able to say for example: "Tree cavity is selected by bats as a roost, when the activity stays in level of XX overfights ("1"'s) for at least XX minutes"

thank you

i will try to do more googling

then maybe a simple logistic regression where time is an IV ?

or something like this

IV?

sorry, this is not precisely my area of expertise

(independent variable)

thank you for your time

:D

bye

try googling or post a Q on CV with describing your problem, it seems interesting and you'd probably get some nice answer by someone who is more proficient in such methods

np, bye

ok, i wil try