VItamin-and-Bilby-with-Different-Noises.md

@@ -54,32 +54,42 @@ Fig.14  random seed = 10000
 We did PE run provided by Bilby twice to see the consistency of the results.    
 ## (1) 4 parameters    
 ### (i) posterior     
+Fig.15  
 ![4](uploads/4149e24039801f12dfaa9d806691a5bf/4.png)    
+Fig.16  
 ![4_2](uploads/b8121b185b496f72825a0db0d7af6dfe/4_2.png)     
 ### (ii) JS divergence     
 #### (a) chirp mass     
 JSD = 0.0498     
+Fig.17  
 ![chirp_mass](uploads/6eb9be9b21d2e25902a13ab54057f622/chirp_mass.png)     
 #### (b) mass ratio       
 JSD = 0.0462     
+Fig.18  
 ![mass_ratio](uploads/930df1dbfaa7af97fdb8c47ec2bd2d62/mass_ratio.png)    
 #### (c) distance     
 JSD = 0.0581     
+Fig.19  
 ![d](uploads/c3c3391f318b691c3214a5995f784ea8/d.png)     
 It can be seen that the consistency of results is good.       
 ## (2) 15 parameters    
 ### (i) posterior    
+Fig.20  
 ![15cbc](uploads/a77d9f9030948fdd57cac91d6d52c0cf/15cbc.png)     
+Fig.21  
 ![15cbc_copy](uploads/a702722d876260267e37fd2e1908b6d4/15cbc_copy.png)      
 ### (ii) JS divergence     
 #### (a) chirp mass     
 JSD = 0.0438    
+Fig.22  
 ![m2](uploads/38d078f7b3c51b860bbed8366d956595/m2.png)     
 #### (b) mass ratio       
 JSD = 0.0472     
+Fig.23  
 ![q2](uploads/7bc25f77785a3baffc213ff923efe293/q2.png)     
 #### (c) distance     
 JSD = 0.0359    
+Fig.24  
 ![d2](uploads/208d8cb343dbe5fd8d9a7f7630a1c3c8/d2.png)     
 It can be seen that the consistency of results is good.         
 Therefore, two tutorial scripts(4 and 15 parameters) can be accepted that they ran well on my laptop.      
@@ -152,49 +162,64 @@ result = bilby.run_sampler(
 ## (3) 15 parameters with (m1,m2) instead of (chirp mass, mass ratio)    
 We get results with good consistency in the sections above. This is to test if (m1,m2) will decrease the convergence.    
 ### (i) posterior    
+Fig.25  
 ![15cbc_n](uploads/3866e0c185dce78896dd1b18100c8a06/15cbc_n.png)     
+Fig.26  
 ![15cbc_n_cp](uploads/6365ee51ec1e0ee52be4b26c5ee0b1e0/15cbc_n_cp.png)     
 ### (ii) JS divergence     
 #### (a) m1     
 JSD = 0.0372     
+Fig.27  
 ![15m1](uploads/0efc51004017f3a3784e5270c8235f9f/15m1.png)    
 #### (b) m2     
 JSD = 0.0459     
+Fig.28  
 ![15m2](uploads/bc181efc370b15debf3f1f8499fd1a24/15m2.png)     
 #### (a) distance     
 JSD = 0.0336       
+Fig.29  
 ![15d](uploads/0e97da3389404c5d4a673e4fb1423e26/15d.png)     
 It can be seen that the consistency of results is still good. Hence we decide to next test how BH encounter performs in the PE run.            
     
 # 4. BH encounter    
 ## (1) BH enoucnter with marginalization(distance, time, phase)     
 ### (i) posterior     
+Fig.30  
 ![encounter](uploads/3681c0bbd9a90c7f5c9684e5ec5a891f/encounter.png)    
+Fig.31  
 ![encounter_cp](uploads/c6fdc02274cdb24f7556690946761aea/encounter_cp.png)     
 ### (ii) JS divergence     
 #### (a) m1     
 JSD = 0.1658     
+Fig.32  
 ![encounter_margil_m1](uploads/aa2d12854aba41d5211441f58c423a02/encounter_margil_m1.png)
 #### (b) m2     
 JSD = 0.1759     
+Fig.33  
 ![encounter_margil_m2](uploads/b28dbf3f5988fa91c87298a66ae2ece2/encounter_margil_m2.png)     
 #### (a) distance     
 JSD = 0.1096    
+Fig.34  
 ![encounter_margil_d](uploads/81863c4571955f98fe5aa3f7d5112dba/encounter_margil_d.png)      
 Wrong template (using BBH template to model BH encounter) reduces the consistency of results.     
 ## (2) BH enoucnter without marginalization(distance, time, phase)     
 ### (i) posterior     
+Fig.35  
 ![encounter_without_margil](uploads/155ebf8256d4bdb4f3aeeab20a974c9b/encounter_without_margil.png)    
+Fig.36  
 ![encounter_without_margil_cp](uploads/b5b138c6f0426eff581a95ed1201de1e/encounter_without_margil_cp.png)     
 ### (ii) JS divergence     
 #### (a) m1     
 JSD = 0.6390     
+Fig.37  
 ![encounter_m1](uploads/190a2c79fccb16ad221cf534fee6ed04/encounter_m1.png)     
 #### (b) m2     
 JSD = 0.4551      
+Fig.38  
 ![encounter_m2](uploads/9d0e47b4162328215a53689855c8454a/encounter_m2.png)     
 #### (a) distance     
 JSD = 0.2937   
+Fig.39  
 ![encounter_d](uploads/1fd6439364cc2a2ba9fa6a061bf778e8/encounter_d.png)     
 Lack of Marginalization of distance, phase and time reduces the consistency of results further.     
 
@@ -205,10 +230,14 @@ https://gilsay.physics.gla.ac.uk/gitlab/wei.changfeng/weichangfeng_project/tree/
 https://gilsay.physics.gla.ac.uk/gitlab/wei.changfeng/weichangfeng_project/tree/master/2021_05/test_2     
 Though they produced the same number of samples this time, I checked the first ten lines of the sample files and they are not the same.      
 ### (i) test 1      
+Fig.40  
 ![test1](uploads/a0d28c07feda2275d7f2862d9ab32a8f/test1.png)      
+Fig.41  
 ![corner1](uploads/5c4fb50dbd6d6a4ade2dfcd78f417ad9/corner1.png)      
 ### (i) test 2           
+Fig.42  
 ![test2](uploads/db55b617fe1a2b00080c0c78f1fe3d27/test2.png)      
+Fig.43  
 ![corner2](uploads/72e14dc47db3e55b95182f4a99b1fc5a/corner2.png)      
 ## (2) Fast tutorial using cpnest sampler.     
 The result still can't be reproduced exactly the same. All the files are uploaded here:     
@@ -216,6 +245,8 @@ https://gilsay.physics.gla.ac.uk/gitlab/wei.changfeng/weichangfeng_project/tree/
 https://gilsay.physics.gla.ac.uk/gitlab/wei.changfeng/weichangfeng_project/tree/master/2021_05/test_cpnest_2      
 They produced the different number of samples. I checked the first ten lines of the sample files and they are not the same.      
 ### (i) test 1      
+Fig.44  
 ![corner_cpn_1](uploads/f1df89d85618e3a74432b06333aad372/corner_cpn_1.png)     
 ### (i) test 2           
+Fig.45  
 ![corner_cpn_2](uploads/81c603af077b79714bfd0dbbf8fcd384/corner_cpn_2.png)     
\ No newline at end of file