murrayefford / secrapp Goto Github PK

This is expected e.g. when lambda0~h2. Problem is to anticipate the shape of the output from predict(). Possible workaround: pre-build newdata and pass to predict; base output on known rows of newdata.

many inputs are ignored

Re-organise and include plot

Requires both estimate of total time/evaluations and re-coding of secr.fit... probably won't happen. However, better modal box warning about time is possible.

Present function timefn is crude and could be put on better basis with simulations and classification of model 'difficulty' (h2 vs rest?)

results dependent on inputs (e.g. HHN/HEX) should be suppressed when the input changes

good to make this available

essentially move 'session' numeric input from Array plot to Captures panel

'use all' option 0?

should be simple to call OU version. Ultimately, pass trap file, density, etc. as parameters

Function modelstring generates a short form of the formulae for the summary. This relies on strsplit to separate the formulae at ',', but this fails when there is a comma internal to a term e.g. s(x,k=6). Use of this function is on hold.

e.g. trap.txt, capt.txt
possibly in /Data subfolder
(details of how to do this all unclear)

prints captdf in some circumstances - not clear why

95%CI is only shown when another box ticked

possibly as separate tabbed page with sections for (i) data input (ii) mask (iii) model fit