Skip to content

Changes

Page history
Update Issues with covaraince matrix authored by Soumen Roy's avatar Soumen Roy
Show whitespace changes
Inline Side-by-side
Issues-with-covaraince-matrix.md
View page @ cbc2dc5d
...@@ -71,7 +71,10 @@ A similar plot for the GW190412 event (top 5 [PE samples](https://git.ligo.org/p ...@@ -71,7 +71,10 @@ A similar plot for the GW190412 event (top 5 [PE samples](https://git.ligo.org/p
The effective-rank of the covariance matrix needs to be tuned for each event using a diagnostic plot such as those shown above. The effective-rank of the covariance matrix needs to be tuned for each event using a diagnostic plot such as those shown above.
## Rank versus $`\beta`$ for off-source samples ## Rank versus $`\beta`$ for off-source samples
<img src="uploads/fe96e535c7f921b3b65f8ba209c47d68/beta_lambda_nalpha_2_.png" width="440" >
<img src="uploads/e37b0ac160bb258219a1b6732dbb2308/beta_lambda_nalpha_4_.png" width="440" >
<img src="uploads/27ac1831fe2b94185cc51bcdba11d53c/beta_lambda_nalpha_3_.png" width="440" > <img src="uploads/27ac1831fe2b94185cc51bcdba11d53c/beta_lambda_nalpha_3_.png" width="440" >
The above figuers show the variation in $`\beta`$ with respect to the rank of the covariance matrix for off-source samples. The plots are produced using the off-source samples obtained for maximum likelihood samples which are reported in the above tables. We can see that the $`\beta`$ values are nearly equal to zero when the rank is very low. But, its modulus value increases sharply within a rank value of 10. The above figuers show the variation in $`\beta`$ with respect to the rank of the covariance matrix for off-source samples. The plots are produced using the off-source samples obtained for maximum likelihood samples which are reported in the above tables. We can see that the $`\beta`$ values are nearly equal to zero when the rank is very low. But, its modulus value increases sharply within a rank value of 10.
......