File: TODO

package info (click to toggle)
r-cran-stabledist 0.7-1-2
  • links: PTS, VCS
  • area: main
  • in suites: bullseye, buster, sid
  • size: 188 kB
  • sloc: makefile: 14
file content (57 lines) | stat: -rw-r--r-- 2,071 bytes parent folder | download | duplicates (2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
* pstable() bug beta ~= -1, found by Anastasija Sanina, Feb.2016
  --->  ~/R/MM/Pkg-ex/stabledist/pstable-uniroot-bug.R

* Compare with Jim Lindsey's "stable" package
  --> http://www.commanster.eu/rcode.html
  Maybe even ask him to make it CRANny
  (maybe after transfering maintainership ?)

  At SfS ETH: source in /usr/local/app/R/R_local/JimLindsey/stable/
  installed in          /usr/local/app/R/R_local/JimLindsey/library

* stableMode():  [MM] I'm pretty sure that the  'beta.max' argument is
  		 unneeded and could be eliminated.

* parametrizations -- pm=0, 1, 2 ---
  provide functions for the parameter transformation that go from
  other to pm=0 -- and back
  and then show how to, or  use these to to provide transformations
  from all pm's to any other.

* [qp] check the  lower.tail & log.p argument settings
  NB: This check *should* happen in  fBasics  'checkdist' ...

* curve(dstable(exp(x), alpha= 0.1, beta=1, pm=1, log=TRUE), -15, 0)
  very much depends on  zeta.tol ... that should somehow depend on alpha or zeta;
  --> tests/dstab-ex.R ("beta = 1")

* As easy alternative to  fBasics::stableFit() define one  based
  on stats4::mle (and possibly also a  quantile-matching alternative).

------------------------------------ Done : --------------------------------------
				     ====

* Nolan ("tail.pdf") and other places mention *asymptotic* formulae
  --> use these for large |x|, notably for  dstable(*, log = TRUE)

* pstable() should also get   lower.tail = TRUE,  log.p = FALSE
  ~~~~~~~   arguments
* qstable() ditto

* using  doExtras to *not* run some tests when  --as-cran ,  as

     From: Prof Brian Ripley <ripley@stats.ox.ac.uk>
     Subject: Re: New version of copula uploaded 0.99-2
     Date: Tue, 29 May 2012 17:17:05 +0100

     ..  but the test times

     * checking tests ...  [173s/174s] OK
	Running doRUnit.R  [82s/83s]
	Running dstab-ex.R [34s/34s]
	Running pstab-ex.R [19s/19s]
	Running tails.R    [39s/39s]

     are more than we would like, so please consider reducing substantially
     next time around
     ...