The activity
table is used to identify batches with too much noise.
Depending on the value of return
these are filtered out, returned, or
identified in the alltags
view with the column probability
. No changes
to the database are made.
Utilisation
filterByActivity(
src,
return = "good",
view = "alltags",
minLen = 3,
maxLen = 5,
maxRuns = 100,
ratio = 0.85
)
Arguments
- src
SQLite connection (result of
tagme(XXX)
orDBI::dbConnect(RSQLite::SQLite(), "XXX.motus")
)- return
Character. One of "good" (return only 'good' runs), "bad" (return only 'bad' runs), "all" (return all runs, but with a new
probability
column which identifies 'bad' (0) and 'good' (1) runs.- view
Character. Which view to use, one of "alltags" (faster) or "alltagsGPS" (with GPS data).
- minLen
Numeric. The minimum run length to allow (equal to or below this, all runs are 'bad')
- maxLen
Numeric. The maximum run length to allow (equal to or above this, all runs are 'good')
- maxRuns
Numeric. The cutoff of number of runs in a batch (see Details)
- ratio
Numeric. The ratio cutoff of runs length 2 to number of runs in a batch (see Details)
Détails
Runs are identified by the following:
All runs with a length >=
maxLen
are GOODAll runs with a length <=
minLen
are BADRuns with a length between
minLen
andmaxLen
are BAD IF both of the following is true:belong to a batch where the number of runs is >=
maxRuns
the ratio of runs with a length of 2 to the number of runs total is >=
ratio
Exemples
# download and access data from project 176 in sql format
# username and password are both "motus.sample"
if (FALSE) sql.motus <- tagme(176, new = TRUE, update = TRUE)
# OR use example sql file included in `motus`
sql.motus <- tagme(176, update = FALSE,
dir = system.file("extdata", package = "motus"))
tbl_good <- filterByActivity(sql.motus)
tbl_bad <- filterByActivity(sql.motus, return = "bad")
tbl_all <- filterByActivity(sql.motus, return = "all")