-
Notifications
You must be signed in to change notification settings - Fork 6
/
NEWS
60 lines (50 loc) · 2.73 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
o Changes in Version 0.3.4
o Minor help file fix
o Changes in Version 0.3.3
o Minor help file fix
o Changes in Version 0.3.2
o Added ebic_tol argument for tolerance when comparing EBIC, if two models now have
near identical EBICs the model with the lowest lambda_kappa value is selected,
followed by the model with the lowest lambda_beta value.
o Changes in Version 0.3.1
o Added a warning for duplicate beeps
o Removed C++11 references
o Changes in version 0.3
o Added arguments regularize_mat_beta and regularize_mat_kappa to only regularize certain edges.
o Fixed a bug in simMLgvar in which the sign of the between-person network was reversed.
o Fixed a bug with the beepvar argument
o Changed deprecated dplyr functions
o Added 'regularize_mat_beta' and 'regularize_mat_kappa' arguments (experimental)
o Changes in version 0.2.3
o Fixed an error due to tibble update
Changes in Version 0.2.2:
o Fixed a CRAN warning
o Added 'lags' and 'likelihood' argments to graphicalVAR
o Fixed message about deprecated functions
Changes in Version 0.2.1:
o Data is now stored in the output of graphicalVAR
o Fixed a bug leading to NA in any column to delete a row
o Added 'likelihood' argument to mimic sparseTSCGM 2.5 behavior
Changes in Version 0.2:
o Added 'tsData' function to prepare data
o graphicalVAR now supports multiple subjects (fixed effects only) and day effects
o 'beepvar' can now be used to handle missing beeps
o Added 'mimic' argument to 'graphicalVAR' to mimic 0.1.2 and 0.1.4 behavior.
o Lambda sequence for kappa now uses the maximum absolute correlation as maximal value rather than 1.
o Added 'mlGraphicalVAR' for pooled and individual estimation of N > 1 datasets
o Added 'simMLgvar' to simulate N > 1 data
o Fixed a bug in 'graphicalVARsim' where mean structure was not used in data generation
Changes in Version 0.1.4:
o 'graphicalVAR' again standardizes variables before running by default. Can be controlled using the 'scale' argument
o New arguments to 'graphicalVAR':
o deleteMissings
o penalize.diagonal
o lambda_min_kappa
o lambda_min_beta
o scale
o Added 'randomGVARmodel' function to simulate graphicalVAR model matrices
o Added unregularized estimation when both lambda_kappa = 0 and lambda_beta = 0
o Greatly updated the tuning parameter sequence generating algorithm. The sequence should now be better chosen. However, note that this change leads to different estimated networks as with previous graphicalVAR versions (as different LASSO tuning parameters are used)
o Added 'lbound' and 'ubound' arguments to graphicalVARsim
Changes in Version 0.1.3:
o graphicalVAR now only centers and does not standardize