fdasrvf 2.3.x

fdasrvf 2.3.3

• bugfixes
• added vervose options throughout

fdasrvf 2.3.2

• bugfixes
• add h representation of warping functions
• added h representation to jfpca

fdasrvf 2.3.1

• fixes for ATLAS BLAS

fdasrvf 2.3.0

• fixed scaling in curve_karcher_mean
• parallelized curve_karcher_mean and curve_srvf_align
• updated sample_shapes and created curveboxplot (#39)
• bugfixes (#38)
• improved distance matrix computation (#45)
• refactor karcher mean (#40 and #47)
• updated README (#50)
• updated plotting routines
• add elastic changepoint functionselastic_amp_change_ff, elastic_ph_change_ff, and elastic_change_fpca

fdasrvf 2.2.0

• bugfixes
• added curve_depth function (#31)
• clarification of multivariate functional data (#31 and #32)
• expanded calc_shape_dist for different pre-shape spaces
• expanded curve functions for different pre-shape spaces
• added predict functions to fpca to project new samples onto basis

fdasrvf 2.1.2

• add rlbfgs c++ code (#30)

fdasrvf 2.1.1

• bugfixes (#27)

fdasrvf 2.1.0

• added elastic change point functions
• exposed SqrtMeanInverse and inv_exp_map to global (#29)
• bugfixes

fdasrvf 2.0.3

• exposed lam to curve functions
• added gamma to shooting vector conversion functions
• bugfixes

fdasrvf 2.0.2

• added dynamic grid to DP method (#25)
• update outputs to curve functions (#26)
• extend exp_map to n-d curves (#27)
• bugfixes

fdasrvf 2.0.1

• Added functionalities to kmeans_align() (#24). Specifically:

  • it gains the argument use_verbose which allows the user to suppress information displayed in the console;
  • it includes a numeric vector storing the distances to corresponding center for each curve in the output;
  • it is now possible to use the medoid as centroid type as an alternative to the mean; the medoid is the most central curve among the existing curves.

• Further small improvements, documentation updates and improved code coverage (#23).

fdasrvf 2.0.0

• Added multivariate kmeans (#18);
• Added additional penalties to warping and additional outputs exposed to bootTB() (#19);
• Switch from travis appveyor to GHA for CI and code coverage monitoring (#20);
• Bugfixes and documentation updates.

fdasrvf 1.9.8

• bugfixes
• remove akima to interp

fdasrvf 1.9.7

• bugfixes

fdasrvf 1.9.6

• bugfixes

fdasrvf 1.9.5

• refactor pf curve functions
• add curve karcher median
• bugfixes

fdasrvf 1.9.4

• add elastic depth functions

fdasrvf 1.9.3

• fix gfortran issues

fdasrvf 1.9.2

• update gropt library

fdasrvf 1.9.1

• Added tolerance bound functions

fdasrvf 1.9.0

• Added elastic principal component regression functions

fdasrvf 1.8.4

• centered rgam calculation
• fix bug in curve functions and image registration

fdasrvf 1.8.3

• added bayesian alignment method by Y. Lu et. al
• added multiple functional alignment function
• bugfixes to curve alignment

fdasrvf 1.8.2

• added objects for outputs of time_warping, boxplot, and pca functions
• added plot methods for the above objects
• added summary methods for the above objects
• added kmeans clustering and alignment based on Sangalli et al.
• multiple bugfixes

fdasrvf 1.8.1

• fix boxplot functions documentation
• added quantile option
• updated surface plot for boxplot functions

fdasrvf 1.8.0

• added phase and amplitude boxplot functions
• fixed issue with time_warping finding the median

fdasrvf 1.7.1

• fix memory leaks in bayesian c code
• update SqrtMean calculation to have more numerical accuracy

fdasrvf 1.7.0

• added a pair align function for 1-D data
• added bayesian alignment for 1-D functions

fdasrvf 1.6.2

• updated curve functions to support R^n
• cleaned up curve function for loops

fdasrvf 1.6.1

• bug fixes to curve functions

fdasrvf 1.6.0

• added open and closed curve function (N-D)

fdasrvf 1.5.2

• added image alignment functions and fix minor bugs

fdasrvf 1.5.1

• added gropt optimization methods and alignment

fdasrvf 1.5.0

• added simul optimization method for reparam

fdasrvf 1.4.2

• Fixed gradient bug

fdasrvf 1.4.1

• Fixed bug in doParallel for windows computers
• Fixed numerical issues for high lambda for no warping required

fdasrvf 1.4

• Fixed memory leak in Dynamic Programming algorithm
• Updated random gamma generation to use a Fourier basis on the tangent space
• Changed over parallel support to doParallel
• Fixed documentation error

fdasrvf 1.3

• Fixed numerical issues
• Updated documentation

fdasrvf 1.2

• Changed SRVF calculation to use splines for derivatives
• Added control to the amount of warping via lambda to the time_warping function
• Fixed documentation bugs and minor bugs in horizontal PCA calculation

fdasrvf 1.1

• Updated so the parallel functions are windows compatible and windows binary can be compiled, now suggests doSNOW or doMC
• Added variance calculation to time_warping function

fdasrvf 1.0

• Initial Version of Package