forked from astroleo/STARLIGHT_fit
-
Notifications
You must be signed in to change notification settings - Fork 0
/
StCv04.C11.config.slow
127 lines (119 loc) · 6.71 KB
/
StCv04.C11.config.slow
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
# Configuration parameters for StarlightChains_v04.for - Cid@Lagoa - 18/Feb/2007 #
#
#
# Normalization lambdas
#
6840.0 [l_norm (A)] = for base spectra only
6810.0 [llow_norm (A)] = for observed spectrum
6870.0 [lupp_norm (A)] = " " "
#
#
# Parameter Limits
#
-1.0 [AV_low (mag)] = lower allowed AV
4.0 [AV_upp (mag)] = upper allowed AV
-0.0001 [YAV_low (mag)] = lower allowed YAV
0.0001 [YAV_upp (mag)] = upper allowed YAV
0.7 [fn_low] = lower allowed Norm. factor = sum x_j
1.3 [fn_upp] = upper allowed Norm. factor = sum x_j
-500.0 [v0_low (km/s)] = lower allowed v0
500.0 [v0_upp (km/s)] = upper allowed v0
0.0 [vd_low (km/s)] = lower allowed vd
500.0 [vd_upp (km/s)] = upper allowed vd
#
#
# Clipping options & Weight-Control-Filter
#
NSIGMA [clip_method_option] = NOCLIP/NSIGMA/RELRES/ABSRES = possible clipping methods
3.0 [sig_clip_threshold] = clip points which deviate > than this # of sigmas
2.0 [wei_nsig_threshold] = weight-control-filter. Use <= 0 to turn this off! (see manual)
#
#
# Miscellaneous
#
50.0 [dl_cushion (A)] = safety margin for kinematical filter!
0.001 [f_cut (units of f_norm)] = Mask/ignore very low fluxes: f_obs <= f_cut
31 [N_int_Gauss] = # of points for integration of kinematical filter
1 [i_verbose] = 0/1 = Quiet/Talkative
0 [i_verbose_anneal] = 0/1/2/3 = Quiet/.../Verborragic
0 [Is1stLineHeader] = 1/0 = Y/N
1 [i_FastBC03_FLAG] = 1 for Fast-rebin of BC03 spectra!
0 [i_FitPowerLaw] = 1/0 = Y/N - include a Power Law in base
-0.5 [alpha_PowerLaw] = PL index, only used if iFitPowerLaw = 1
0 [i_SkipExistingOutFiles] = 1/0 = Y/N - skip or overwrite fits with already existent arq_out
#
#
# Markov Chains technical parameters
#
12 [N_chains] = # of Markov Chains
0.50 [xinit_max] = max(x_j) for initial random chain pop-vecs
0 [i_UpdateEps] = 1/0 = Y/N. Not well tested: use 0!
2 [i_UpdateAlpha] = 0/1/2. 1 & 2 update step-sizes dynamically. 0 turns this off.
2.0 [Falpha] = step-updating-factor.
1 [i_MoveOneParOnly] = 1/0 = Y/N. Not tested/debugged! Use 1!
1 [i_UpdateAVYAVStepSeparately] = 1/0 = Y/N.
1 [i_HelpParWithMaxR] = 1/0 = Y/N. Help convergence of ParWithMaxR
0.2 [prob_jRmax] = prob to pick ParWithMaxR
1 [i_HelpPopVectorMove2Average] = 1/0 = Y/N. Help x convergence
0.4 [prob_HelpPopVectorMove2Average] = prob of inverting sign of x-move to go towards mean
1 [i_HelpAVYAVMove2Average] = 1/0 = Y/N. Help AV/YAV convergence
0.4 [prob_HelpAVYAVMove2Average] = prob of inverting sign of AV/YAV-move to go towards mean
10 [NRC_AV_Default] = initial # of terms in the RC AV-series
#
#
# First Fits (FF) technical parameters
#
1.0e2 [Temp_ini_FF] = initial Temperature
1.0 [Temp_fin_FF] = final Temperature
1.0e1 [fini_eps_FF] = initial step-size (fraction of low->upp range)
1.0e2 [ffin_eps_FF] = final step-size (fraction of low->upp range)
1.1 [R_ini_FF] = initial GR-R convergence threshold
1.1 [R_fin_FF] = final GR-R convergence threshold
0 [IsGRTestHard_FF] = 0/1 = Hard/Soft GR-R convergence criterion
10 [N_loops_FF] = # of annealing loops
1 [i_RestartChains_FF] = 1/0 = Y/N = Reset chains to best pars after each loop
1.0e1 [fN_sim_FF] = x N_par = length of each chain loop...
1.0e4 [fNmax_steps_FF] = max length of each chain loop...
0.23 [eff_IDEAL_FF] = ideal efficiency, used to optimize step-sizes
#
#
# GR R-threshold & Method for Burn-In loop
#
1.2 [R_Burn_in] = GR-R convergence threshold for burn-in
1 [IsGRTestHard_BurnIn] = 0/1 = Hard/Soft GR-R convergence criterion
#
#
# EX0s technical parameters
#
MIN [EXOs_PopVector_option] = MIN/AVE = Use x_min or x_ave to define condensed base
CUMUL [EXOs_method_option] = CUMUL/SMALL = method to define irrelevant components
0.02 [EXOs_Threshold] = Irrelevance threshold
1.0 [Temp_ini_EX0s] = Analogous to FF entries above!
1.0e-3 [Temp_fin_EX0s] = "
1.0e2 [fini_eps_EX0s] = "
1.0e3 [ffin_eps_EX0s] = "
1.1 [R_ini_EX0s] = "
1.1 [R_fin_EX0s] = "
1 [IsGRTestHard_EX0s] = "
10 [N_loops_EX0s] = "
1 [i_RestartChains_EX0s] = "
1.0e2 [fN_sim_EX0s] = "
1.0e3 [fNmax_steps_EX0s] = "
0.50 [eff_IDEAL_EX0s] = "
1 [IsScaleNstepsInEX0sFits] = 1/0 = Y/N = Scale numbers of steps by # of EX0 components.
1 [IsNoKin4LargeBaseInEX0sFits] = 1/0 = Y/N = Stop v0 & vd fits in EX0s for large bases
0.0 [frac_NoKin4LargeBaseInEX0sFits] = Will stop v0 & vd fits if EX0-bases > frac_... * N_base
0.1 [fEX0_MinBaseSize] = EX0-fits will use at least fEX0_MinBaseSize * N_base components
Cid@Lagoa - 18/February/2007
OBS: This config is a reformatted (for v04) & renamed version of the
StCv03.t02.config used in the config tests of Jan/2007.
Technical parameters you may want to play with to obtain FAST, MEDIUM
& SLOW fits:
--------------------------------
| FAST | MEDIUM | LONG |
--------------------------------
| 5 | 7 | 12 | [N_chains]
| 3 | 5 | 10 | [N_loops_FF & *_EX0s]
| 1.3 | 1.2 | 1.1 | [R_ini_FF & R_fin_FF & *_EX0s]
| 0 | 0 or 1 | 1 | [IsGRTestHard_FF & *_BurnIn & *_EX0s]
--------------------------------