1Changes in Version 1.5-2 2 3 o Added lrvar = FALSE argument to efp() so that optionally a long-run variance 4 estimator (Andrews or Newey-West) instead of the standard OLS estimator can 5 be used for the error variance. 6 7 o Added vcov. = NULL argument to efp() so that optionally other covariance 8 matrix estimators can be plugged into RE and ME tests (e.g., sandwich(), 9 kernHAC(), NeweyWest(), etc.). 10 11 o The default recresid() gained an argument qr.tol = 1e-7 which allows to 12 pass another tolerance to lm.fit() for detecting linear dependencies in 13 (small) subsamples. Furthermore, an argument engine = c("R", "C") has been 14 added along with an alternative faster C implementation. 15 16 o The formula method of breakpoints() now passes ... to recresid(), e.g., for 17 the qr.tol and engine arguments above. 18 19 o The breaks argument of the breakpoints() formula argument is now checked 20 to be at least 1. 21 22 o breakpoints(..., hpc = "foreach") now also works if the "foreach" package 23 is not attached. 24 25 o Improved support for formulas like 'y ~ .' in efp(), Fstats(), and 26 breakpoints(). (Suggested by Matthieu Stigler.) 27 28 o Bug fix in gefp(..., decorrelate = FALSE). Scaling is done with the square 29 root of the diagnoal of the variance - as opposed to the diagonal of the 30 square root of the variance. (Reported by Dries Debeer.) 31 32 33Changes in Version 1.5-1 34 35 o ordL2BB() now uses a direct simulation method based on mvtnorm::rmvnorm() 36 which is much faster, making the computation of p values and critical values 37 for the ordinal maxLM statistic much faster and feasible "on the fly". 38 39 o Reduced number of significant digits in breakpoints summary to 40 getOption("digits") - 3. 41 42 o Reference output updated for recent versions of R. 43 44 45Changes in Version 1.5-0 46 47 o Added new efpFunctional generators for conducting various types of 48 structural change tests based on empirical fluctuation processes of class 49 "gefp". In particular a (maximum) MOSUM functional was added as well 50 as several functionals suitable for aggregation along categorical 51 variables. The documentation for previously available functionals 52 such as supLM() was also enhanced. 53 54 o The new functionals mentioned above for assessing parameter instability 55 along (ordered) categorical variables are catL2BB (unordered), 56 ordL2BB and ordwmax (ordered). These are discussed in more detail in 57 Merkle, Fan, and Zeileis (2013, Psychometrika). 58 59 o Added a new default method for sctest(). This essentially just calls 60 gefp(object, fit = NULL) and then (optionally) calls plot() and sctest() 61 using the specified functional. However, several convenience options have 62 been added, e.g., using the maximum likelihood information (rather than 63 the outer product of gradients) for the covariance matrix or specifying 64 the functional via a character string. 65 66 o Documentation of the sctest() generic and its methods have been enhanced. 67 Methods for "formula", "efp", and "Fstats" are suitable for assessing 68 structural changes in linear regression models while the "default" and 69 "gefp" methods (see above) are suitable for general parametric models. 70 71 o Improved plot() method for gefp/efpFunctional to allow for more 72 flexibility in boundary display. Rather than only boundary = TRUE) or 73 FALSE one can now specify a list of graphical parameters, e.g., 74 boundary = list(col = "slategray", lty = 2). 75 76 o Updated Depends/Imports in DESCRIPTION/NAMESPACE with new R CMD check 77 requirements. 78 79 80Changes in Version 1.4-7 81 82 o plotProcess() function in "efpFunctional" objects now takes a 83 boundary = TRUE argument by default which can be set to FALSE to 84 suppress plotting of the boundary function. 85 86 o Added a check (and a more intelligible warning) in the "formula" 87 method of breakpoints() whether the 'breaks' argument supplied by 88 the user is too large. 89 90 91Changes in Version 1.4-6 92 93 o Default recresid() can now also deal with regressors that do not 94 vary across (small) subsamples. 95 96 97Changes in Version 1.4-5 98 99 o Further improvements in new recresid() default method. 100 Now also works correctly if some coefficients are not identified 101 on the initial subsamples in the recursion. 102 103 o Resaved datasets to reduce storage requirements. 104 105 o Fixed bug in breakpoints() for time series that contain NAs. 106 107 108Changes in Version 1.4-4 109 110 o Default recresid() method now tries to choose adaptively between 111 using the faster updating formula and the slower full 112 QR decomposition to yield numerically more stable results. 113 In previous versions of the function the QR decomposition was 114 used only in the first iteration. 115 116 o Improvement in breakdates() computations. 117 118 119Changes in Version 1.4-3 120 121 o Speed-up in breakpoints() for the intercept-only case, 122 i.e., breakpoints(y ~ 1). 123 124 125Changes in Version 1.4-2 126 127 o Improved time index computations in gefp(). 128 129 o Added replication notes in ?durab. 130 131 132Changes in Version 1.4-1 133 134 o efp(), Fstats(), and breakpoints() are now more cautios about using 135 time series properties from the data and try to check whether any 136 NAs were removed. In general, the functions will yield best results 137 if all NA processing is done before calling them. 138 139 o Better handling of time series properties for the boundaries in 140 the examples of SP2001. 141 142 143Changes in Version 1.4-0 144 145 o added optional high performance computing support by means of the 146 "foreach" package for the breakpoints() formula method. This can 147 be leveraged to alleviate the computational burden in the dynamic 148 programming approach. Simply register a parallel backend (e.g., 149 by means of "doMC" or "doSNOW") and call breakpoints() with 150 additional argument hpc = "foreach". 151 152 153Changes in Version 1.3-7 154 155 o added optional start end end arguments to recresid(). 156 157 158Changes in Version 1.3-6 159 160 o enhanced documentation for new Rd parser. 161 162 163Changes in Version 1.3-5 164 165 o added some further references to the vignette, 166 and provide the associated .bib file in ~/inst/doc/. 167 168 o removed \itemize in .Rd files for new R-devel 169 170 171Changes in Version 1.3-4 172 173 o fixed CITATION encoding 174 175 o removed Z.sty dependency in vignette 176 177 178Changes in Version 1.3-3 179 180 o enhanced references in the vignette, CITATION and 181 man pages 182 183 o fixed some outdated information in the vignette 184 185 186Changes in Version 1.3-2 187 188 o added new data set with bibliographic information about 189 structural change publications 190 191 192Changes in Version 1.3-1 193 194 o renamed SP500 to SP2001 to avoid conflicts with MASS 195 196 197Changes in Version 1.3-0 198 199 o added NAMESPACE 200 201 o improved dependency declaration in DESCRIPTION 202