forked from pbreheny/ncvreg
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathNEWS
159 lines (143 loc) · 6.73 KB
/
NEWS
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
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
3.6-0
* Improved efficiency for cox models (avoids recalculating linear
predictors)
* Exported std() function for standardizing a design matrix
* Reorganized testing suite
* Fixed bug in predict.cv.ncvsurv
* Added 'quick start' vignette
* 'survival' package now used for setupLambda in Cox models
3.5-2
* Fixed bug for ncvsurv with integer penalty factors
* Fixed rare numerical accuracy bug in cv fold assignments
* Fixed LOOCV bug introduced by bias-correction feature
* Added user interrupt checking
3.5-1
* Fixed bug with penalty.factor for cv.ncvsurv when some columns may be
degenerate
* Replaced AUC function with more efficient version using survival
package
* Compute bias correction for CV error; this is an experimental feature
at this point and may change in the future
3.5-0
* Substantial efficiency improvements throughout for Cox models.
Coordinate descent redesigned to work in O(n) instead of O(n^2)
operations, and R code redesigned at various points to avoid the
creation of any n x n matrices when fitting and cross-validating Cox
regression models.
* Added function AUC() to calculate cross-validated AUC values for
ncvsurv models.
* Changed method for calculation of cross-validation loss in
cv.ncvsurv().
* More accurate calculation for convexMin in the presence of unpenalized
variables
* Bug fix for factor-valued y with CV logistic regression
* Added option to return fitted values from cross-validation folds
(returnY=TRUE) to cv.ncvreg and cv.ncvsurv.
* Better double/int type checking for penalty.factor
* Modifications to NAMESPACE for compatibility with R 3.3.
3.4-0
* Expanded predict function for Cox models. predict.ncvsurv now
estimates subject-specific survival functions and medians.
* Added plot method for survival curves.
* Added option in perm.ncvreg to permute residuals for linear regression
* On a related note, added permres() function to estimate false
inclusion rates based on residuals at a specific value of lambda
* Added some support for factors in X, y. It is still recommended that
users convert X to a numeric matrix prior to fitting in order to
ensure that predict() methods work properly, but ncvreg will now allow
you to pass a data frame with factors and handle things appropriately.
* Fixed a bug for predict.ncvsurv when applied to models with
saturation issues
* Fixed small memory leak in ncvsurv
3.3-0
* Support for fitting survival models added (ncvsurv), along
with predict, plot, and cv.ncvsurv support functions. Currently,
Cox models are the only type of survival model implemented.
* Added parallel support for cv.ncvreg (with help from Grant Brown)
* Fixed bug in cv.ncvreg when attempting to use leave-one-out
cross-validation (thank you to Cajo ter Braak for pointing this
out)
* ncvreg_fit taken offline; it will return in a future version of
the ncvreg package.
3.2-0
* Updated algorithm to 'hybrid' strong/active cycling
* Automatically coerces X to matrix and y to numeric if possible
* Made ncvreg_fit more user-friendly: user no longer has to
specify lambda, works with coef, predict, plot, etc.
* Fixed bug in convexMin when used with penalty.factor option
* Modified order of arguments for predict so that 'type' comes
before 'lambda' and 'which'.
3.1-0
* Added support for Poisson regression
* Fixed bug in ncvreg_fit that could arise when fitting a model
without an intercept
* Fixed bug in cv.ncvreg with univariate regression (thank you to
Diego Franco Saldana for pointing this out)
3.0-0
* Revised internal algorithms to incorporate targeted cycling
based on strong rules
* Moved standardization to C
* Moved calculation of lambda sequence to C
* As a result of the above three changes, ncvreg now runs
much faster for large p
* Added fir(), perm.ncvreg(), and plot.fir() functions for the
purposes of estimating and displaying false inclusion rates;
these are likely to evolve over the next few months
* Fixed a bug in cv.ncvreg for user-specified lambda sequence
2.7-0
* Added "vars" and "nvars" options to predict function.
* Modified look of summary(cvfit) output.
* Modified internal details of .Call interface.
2.6-0
* Introduction of function ncvreg_fit for programmers who want to
access the internal C routines of ncvreg, bypassing internal
standardization and processing
* Internal restructuring: .Call now used instead of .C
* Bug fix for axis annotations with plot.cv.ncvreg when model is
saturated.
* Bug fix for deviance calculation; would return NaN if fitted
probabilities of 0 or 1 occurred for binomial outcomes.
* Bug fix in NAMESPACE for coef.cv.ncvreg and predict.cv.ncvreg
* Added vertical.line and col options to plot.cv.ncvreg
2.5-0
* Added options in plot.cv.ncvreg to plot estimates of r-squared,
signal-to-noise ratio, scale parameter, and prediction error in
addition to cross-validation error (deviance)
* Added summary method for cv.ncvreg which displays the above
information at lambda.min, the value of lambda minimizing the
cross-validation error.
* Fixed bug in cv.ncvreg with user-defined lambda values.
2.4-0
* Fixed error in definition/calculation of cross-validation
error and standard error
* Added penalty.factor option
* cv.grpreg: Now returns full data fit as well as CV errors
* coef and predict methods now accept lambda as argument
* Added logLik method (which in turn allows AIC/BIC)
* Fixed bug that arose if lambda was scalar (instead of a vector)
* cdfit_ now returns loss (RSS for gaussian, deviance for binomial)
* Fixed bug in cv.ncvreg for linear regression -- cross-validation
was being carried out deterministically
(Thank you to Brenton Kenkel for pointing this out)
* Internal change: standardization more efficient
* Fixed bug: Intercept for logistic regression was not being
calculated for lamda=0
2.3-2
* Fixed formatting error in citation.
2.3-1
* plot.ncvreg: Made the passing of arguments more flexible, so
that user can pass options concerning both the plot and the lines.
* plot.ncvreg: Changed some of the default settings with respect
to color (hcl instead of hsv) and line width.
2.3
* cv.ncvreg.Rd: Fixed the documentation, which no longer agreed
with the function usage. This was an oversight in the release of
version 2.2.
2.2
* cv.ncvreg: Divorced cross-validation from fitting. From a user
perspective, this increases flexibility, although obtaining the
model with CV-chosen regularization parameter now requires two
calls (to ncvreg and cv.ncvreg). The functions, however, are
logically separate and involve entirely separate methods.
* plot.cv.ncvreg: Developed a plotting method specific to
cv.ncvreg objects.