forked from bjbiggerstaff/PooledInfRate
-
Notifications
You must be signed in to change notification settings - Fork 3
/
CHANGELOG
59 lines (36 loc) · 2.75 KB
/
CHANGELOG
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
58
59
Version 1.6
-----------
User visible:
1. Fixed an error in the naming of the object returned by vectorIndex and VI when there is only one group.
Version 1.5
-----------
User visible:
1. Revised the upper limit for the score interval to match the Wilson interval for the unpooled case when all pools are negative (for a perfect test only).
Version 1.4
-----------
User visible:
1. Added as.data.frame.* methods for pooledBin/pIR, pooledBinDiff/pIRDiff, ipooledBin, vectorIndex/VIR.
Version 1.3
-----------
User visible:
1. Fixed a bug in printing of the default calls to pooledBin/pIR.
Version 1.2
-----------
User visible:
1. Added facility to include multiple grouping variables to the formula interface (but not to the default interface) for pooledBin/pIR, vectorIndex/VI, and ipooledBin functions; this uses an asterisk (*) to separate the groups after the "|" in the formula, such as G1*G2 for groups G1 and G2. (Note that as of now this is not available for the difference functions pooledBinDiff/pIRDiff. These still require just one grouping variable at a time, since all pairwise comparisons within the group are computed and it is unclear which comparisons may be wanted with multiple groups.)
2. Fixed output name "PointEst" to "P" for summary calls to pooledBin/pIR.
Version 1.1
-----------
User visible:
1. Removed records with n=0 or m=0 in pooledBin/pIR, pooledBinDiff/pIRDiff, and vectorIndex/VI rather than the call resulting in what appears to be a convergence error.
2. Changed the basic returned objects for all classes to (essentially) data frames for ready and more intuitive assignment and subsetting, as with data frames. These objects still carry attributes, accessible with attr(obj,"<attribute name>").
a. Because returned objects were not of class data frame, this meant including "subsetting" ("[") methods for all the classes.
b. These functions are really just copies of the built-in code for [.data.frame, though because .Internal() calls cannot be included in packages, a slight revision was needed at the single line this occured; this is noted in the code.
3. Added an argument simple = TRUE/FALSE to summary methods for pooledBin/pIR and pooledBinDiff/pIRDiff to return the detailed result tables as data frames.
Behind the scenes:
1. Removed the "List" structure for dealing with more than one group
a. In the original code design, separate functionality was to keep the case with just one group separate for a different printing style (historical consistency). In hindsight, this was an unnecessary and inefficient complication, so this update presents consistent output for one or more groups.
b. This resulted in a great "slimming down" of the code.
Version 1.0
-----------
1. Initial release