2023-9-14 Marcello Chiodi
*fortran_etas_code.f90: all INTEGER(KIND=4) changed to INTEGER
2023-7-18 Marcello Chiodi
*update.etasclass.R: 'if (class(object) != "etasclass")' changed to 'if (!inherits(object,"etasclass"))' changed also in other R files whenever 'if (class(object) != "etasclass")' was present
2023-7-17 Marcello Chiodi
*fortran_etas_code.f90: arcaic form derfc changed to erfc
2021-9-1 Marcello Chiodi
*etasclass.R: Changed printing during output
2021-8-24 Marcello Chiodi
*etasclass.R: variable magn1.work has been renamed as magnitude
*b.guten: a bug has been corrected (thanks to Jiancang Zhuang)
original first line "x = subset(magn, subset = (magn > m0))"
corrected to "x = subset(magn, subset = (magn >= m0))"
2021-8-1 Marcello Chiodi
*etas.starting.R: a bug has been corrected (thanks to Gordon Ross), for the case when longlat.to.km is FALSE
2021-8-1 Marcello Chiodi
*etasclass.R: a bug has been corrected (thanks to Gordon Ross), and inferior limit for p is now 1.
2021-6-19 Marcello Chiodi
*timeupdate.etasclass.R: A function to update etasclass objects for new observations has been added. Options can be given and they will override previous values
*update.etasclass.R: An update method for etasclass objects has been added. A new iteration will run starting from the full output of the previous object. Options can be given and they will override previous values
*etasclass The algorythm has been partially rewritten and the sequence of optimization, density estimation and weighting has a different order
*etasclass: a new argument n.iterweight has been added: the weighting and the density computations will be alternated n.iterweight times after each maximum likelihood step: in many situations this improves the general convergence procedure
*etasclass: In the output all values of input arguments are given either user values or default values: this is useful for the new update method
2021-1-13 Marcello Chiodi
*print.etasclass: some fields added in the printing of etasclass objects
*daily.etasclass: a new function added to compute daily maps of predictions
*etasclass: many internal changes in computing and managing of starting points. Optional argument cat.back avoided from this version; default value for sectoday changed in FALSE; call to etas.starting to obtain starting values for mu and k0; catalog fields changed and extended in final output; rho.weights updated only after the ML step; default bandwidth computed;
n.iterweight is a new optional argument, which fix the number of iterations made in recomputing weights
*etas.starting default value for sectoday changed in FALSE; argument m0 renamed with magn.threshold for coherence with etasclass; argument onlytime introduced for compatibility with etasclass
2020-7-17 Marcello Chiodi
*data: earthquakes catalogs has been changed in compressed format
*citation: a new reference of year 2020 added for the ETAS model with covariates
2020-7-15 Marcello Chiodi
*etasFLP-package: a new reference of year 2020 added for the ETAS model with covariates
*plot.etasclass: two bugs fixed
*NAMESPACE: dependence on package fields added
2020-4-22 Marcello Chiodi
*etasFLP: dropped out the parallel option to avoid problems on some compiler.
*etasclass: removed the options fastML,ellipse.trig,rxy also in all functions
2019-02-28 Marcello Chiodi
*etasclass: a options introduced parallel.density e parallel.etas, passed to interfaces for FORTRAN routines
*fortran_etas_code.f90: parallel versions of etasfull8NEW and of density2 have been introduced
*Makevars: added support for openMP compiler for parallel computing
2019-01-28 Marcello Chiodi
*etasclass: some improvement introduced
2018-02-28 Marcello Chiodi
*kde2dnew.fortran: two new arguments added hvarx, hvary, vectors of bandwidth
(to be documented)
*etasclass: two new arguments, formula1 and offset has been added to allow for introduction of covariates
*etasclass: the output object of the class etasclass, has been largely modified
*etasclassNEW: not used
*etasclass: a new function generaloptimizationNEW has been introduced, which choose the method of optimization
2018-01-31 Marcello Chiodi
*etasclass: a new argument fastML has been added
*etasclass: ellipse.trig and rxy for ellipsoidal spatial function in triggered intensity (not yet completed)
2018-01-31 Marcello Chiodi
*etasclass: a new argument fastML has been added
2017-05-26 Marcello Chiodi
*etasclass: a new argument, formula1 has been added to allow for future introduction of covariates
*etasclass: the output object of the class etasclass, has been extended
*etasclass: x.km,y.km variables renamed xback.work,yback.work
*etasclass: xcat.km,ycat.km variables renamed xcat.work, ycat.work
*etasclass: a call to a new subroutine generaloptimization added in june 2017 as an interface to nlm or optim according to the option usenlm.
*etas.mod2: process parametrs pass though a unique argument etas.obj of class "etasclass".
*etasclass: inserted a new argument ellipse.trig, that allows for ellipsoidal space triggered intensities.
*weighted.space: a new subroutine used when ellipse.trig=TRUE; essentially an interface to FORTRAN subroutine wspace.
*cat.select: a new subroutine used to select the catalog on which etasclass will work.
2017-04-01 Marcello Chiodi
*R package etasFLP: changes from version 1.4.0 to version 1.4.1
*general changes: the CITATION file quotes a reference to a registered JSS (Journal of Statistical Software) article.
*etasclass, etas.mod2: inserted a new argument tmax, that can be different from the observed maximum.
*plot.etasclass: Some plot label have been changed.
2017-01-04 Marcello Chiodi
*R package etasFLP: changes from version 1.3.1 to version 1.4.0
*general changes: added a CITATION file with the reference to the JSS (Journal of Statistical Software) article (the DOI in the CITATION is for a new JSS publication that will be registered after publication on CRAN)
*plot.etasclass: added tfixed and totfixed in the output.
2016-12-07 Marcello Chiodi
*R package etasFLP: changes from version 1.3.0 to version 1.3.1
*general changes: used different Fortran compiler options to avoid some warnings in the CRAN Package Check Results page
2015-04-09 Marcello Chiodi
*R package etasFLP: changes from version 1.2.1 to version 1.3.0
*general changes: bug corrected in flp computation. Results obtained with previous versions can have little differences with the present ones.
*etasflp.integral: function etasflp.integral has been dropped out; it is useless in the flp optimization step.
*etasclass.Rd, summary.etasclass.Rd, profile.etasclass.Rd: changed the magnitude threshold of the example, so that the whole data of "italycatalog" are used.
*R package etasFLP: changes from version 1.2.0 to version 1.2.1
*etas.starting: corrected a bug when using geographical coordinates (the routine is still a beta-version)
*etasclass: now as input catalog an object of class eqcat or data.frame can be given. Inserted a better choice of the starting values of the bandwidth passed to the function flp1.etas.nlm in the flp step
2014-12-22 Marcello Chiodi
*R package etasFLP: changes from version 1.1.1 to version 1.2.0
*etas.starting: new function to estimate starting values (beta-version)
*plot.etasclass: added several plot for diagnostic and residual analysis (ten plots in four windows)
*plot.etasclass: Corrected a bug for the computation of integrated triggered intensities
*plot.etasclass: Checked a bug when "backconstant=TRUE"
*simpson.kD: the function now is exported for end users
2014-10-30 Marcello Chiodi
*R package etasFLP: changes from version 1.1.0 to version 1.1.1
*summary.etasclass.R: Corrected a bug in printing the output
2014-10-27 Marcello Chiodi
*R package etasFLP: changes from version 1.0.3 to version 1.1.0
*compare.etasclass: added the function "compare.etasclass" to compare two "etasclass" objects
*etasflp.integral: a bug has been fixed
*plot.etasclass: changed some default values. Added the optional argument "tfixed"; if specified, the triggered intensity estimated at day "tfixed" is plotted
*etasclass: changed some default values. Added the optional argument "longlat.to.km": if "TRUE" the long and lat are considered geographical coordinates; input initial values for "hdef" (bandwidths) and "w" (weights)
can be given;
2014-04-14 Marcello Chiodi
*R package etasFLP: changes from version 1.0.2 to version 1.0.3
*NAMESPACE: improved the import export instructions; S3 methods explicitly registered. Only few functions of the package are explicitely exported.
*etasclass: improved the printing of some execution message, avoiding "print()".
2014-03-27 Marcello Chiodi
*R package etasFLP: changes from version 1.0.1 to version 1.0.2
*etasclass.R: Corrected two bugs (when the hessian has a determinant close to zero, and when onlytime =TRUE); added the argument "method", used only if "usenlm=FALSE": method used by "optim"; the use of the starting background catalog ("cat.back") has been improved.
*plot.etasclass: Corrected a bug in the computation of back.grid.
*magn.plot: The argument "..." has been added, so that graphical options can be passed to customize graphical output.
*etasFLP-manual.pdf: Some correction has been made in the documentation files.
2014-02-10 Marcello Chiodi
*R package etasFLP: changes from version 1.0.0 to version 1.0.1
*print.etasclass.R: function added. S3 method for
the generic function print for objects of class "etasclass".
*print.etasclass.Rd: added documentation file for the new function "print.etasclass.R".
*californiacatalog.RData: a new data set added to the package with a North California earthquake catalog. Documented in the help file "californiacatalog.Rd".
*californiacatalog.Rd: added documentation file for the new data set "californiacatalog.RData".
*DESCRIPTION file: Avoided dependency on package "car"; external packages "fields", "rgl", "maps", "mapdata" declared in "Imports" row instead of "Depends" row, due to a note present on the CRAN test page.
*plot.profile.R: small changes to the titles of the plots.
*summary.etasclass.R: small changes to the output format.
*etasclass.R: the input argument params.ind now is is a logical vector (default is the same: replicate(8,TRUE)); instead of a vector of 0 and 1; the meaning is the same.
Corrected a bug in the comparison between the successive declustering iterations
In the output the element "this.call" has been added and contains the call of the function.
Corrected and improved some headers in the names of the output elements
*plot.etasclass: added two new input arguments: "PDF" and "file"; if "PDF=TRUE" then the output is sent to pdf file with name "file".
Some plot label and titles have been corrected.
*profile.etasclass.Rd: correction in the item nprofile.