Difference between revisions of "Error propagation in Amplitude Analysis"

From UConn PAN
Jump to navigation Jump to search
Line 38: Line 38:
 
= \sum_{\gamma,\delta}{\rho_{\gamma\delta}
 
= \sum_{\gamma,\delta}{\rho_{\gamma\delta}
 
     \sum_{\alpha,\beta}^n{
 
     \sum_{\alpha,\beta}^n{
−
       u_\alpha u_\beta^* J_{\alpha \beta}
+
       u_\alpha u_\beta^* J^{\gamma\delta}_{\alpha \beta}
 
     }
 
     }
 
   }
 
   }
Line 46: Line 46:
  
 
When considering the uncertainty on the overall integral, both the errors on ''u'' parameters  
 
When considering the uncertainty on the overall integral, both the errors on ''u'' parameters  
−
and those from the finite MC set of events will contribute. A single detected event (i) can be viewed as one sample of a Poisson process, having therefore an uncertainty of  
+
and those from the finite MC set of events will contribute. A single detected event (i) can be viewed as one sample in a process of independent events, having therefore a count uncertainty of  
 
&sigma;<sub>i</sub>=1. An integral over such events is then a weighted sum of such samples,  
 
&sigma;<sub>i</sub>=1. An integral over such events is then a weighted sum of such samples,  
−
having therefore a contribution to the variance:
+
having resulting in a contribution to the variance:
  
 
<math>
 
<math>
Line 79: Line 79:
 
<math>
 
<math>
 
\sigma_{fit}^2=
 
\sigma_{fit}^2=
−
\sum_{k,l}^n{ \sigma_{u_k}\sigma_{u_l}
+
\left| \sum_k^n{ \sigma_{u_k}
 
   \frac{\partial}{\partial u_k}\left(
 
   \frac{\partial}{\partial u_k}\left(
−
   \sum_{\gamma,\delta}{\rho_{\gamma\delta}
+
   \sum_{\gamma,\delta}{ \rho_{\gamma\delta}
−
     \sum_{\alpha,\beta}^n{
+
     \sum_{\alpha,\beta}^n{u_\alpha u_\beta^* J_{\alpha\beta}}
−
       u_\alpha u_\beta^* J_{\alpha\beta}
+
  }
 +
  \right)
 +
}
 +
\right|^2
 +
</math>
 +
::<math>
 +
=\left( \sum_k^n{ \sigma_{u_k}
 +
  \sum_{\gamma,\delta}{ \rho_{\gamma\delta}
 +
    \sum_{\alpha,\beta}^n{\delta_{k\alpha} u_\beta^* J^{\gamma\delta}_{\alpha\beta}}
 +
  }
 +
}\right)
 +
\left( \sum_{k'}^n{ \sigma^*_{u_{k'}}
 +
  \sum_{\gamma',\delta'}{ \rho_{\gamma'\delta'}
 +
    \sum_{\alpha',\beta'}^n{\delta_{k'\alpha'} u_{\beta'} J^{\gamma'\delta'*}_{\alpha'\beta'}}
 +
  }
 +
}\right)
 +
</math>
 +
::<math>
 +
= \sum_{\gamma,\delta,\gamma',\delta'}{
 +
    \rho_{\gamma\delta} \rho_{\gamma'\delta'}
 +
    \sum_{\alpha,\beta,\alpha',\beta'}^n{
 +
       \left(\sigma_{u_\alpha}\sigma^*_{u_{\alpha'}}\right)
 +
      \left(u_\beta^* J^{\gamma\delta}_{\alpha\beta}\right)
 +
      \left(u_{\beta'} J^{\gamma'\delta'*}_{\alpha'\beta'} \right)
 
     }
 
     }
 
   }
 
   }
−
  \right)
+
</math>
−
  \frac{\partial}{\partial u_l}\left(
+
::<math>
−
  \sum_{\gamma',\delta'}{\rho_{\gamma'\delta'}
+
= \sum_{\gamma,\delta,\gamma',\delta'}{  
−
     \sum_{\alpha',\beta'}^n{
+
    \rho_{\gamma\delta} \rho_{\gamma'\delta'}
−
       u_{\alpha'} u_{\beta'}^* J_{\alpha'\beta'}
+
     \sum_{\alpha,\alpha'}^n{
 +
       \left(\sigma_{u_\alpha}\sigma^*_{u_{\alpha'}}\right)
 +
      G^{\gamma\delta*}_\alpha G_{\alpha'}^{\gamma\delta}
 
     }
 
     }
 
   }
 
   }
−
  \right)
 
−
}
 
 
</math>
 
</math>
−
The product of &sigma; terms in the summation are the error matrix derived from the fit.
+
 
 +
The product of &sigma; terms in the summation is represented by the error matrix derived from the fit. ''G'' was defined as
 +
 
 +
<math>
 +
G_\alpha^{\gamma\delta}=\sum_\beta{ u_\beta J_{\alpha\beta}^{\gamma\delta}}
 +
</math>
 +
 
 +
The overall uncertainty in the integral ''I'' defined in the beginning comes out to:
 +
 
 +
<math>
 +
\sigma_I=\sqrt{\sigma^2_{MC} + \sigma^2_{fit}}
 +
</math>

Revision as of 03:21, 22 November 2011

The following is a review of error propagation needed in amplitude analysis.

Consider a Monte-Carlo (MC) integral over the intensities of N detected events out of Ngen generated.

Failed to parse (MathML with SVG or PNG fallback (recommended for modern browsers and accessibility tools): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle I=\frac{1}{N_{gen}}\sum_i^N{ \sum_{\gamma,\delta}{\rho_{\gamma\delta} \left| \sum_\alpha^n{ u_\alpha A_\alpha^{\gamma \delta}(x_i) } \right|^2 } }= \frac{1}{N_{gen}}\sum_i^N{ \sum_{\gamma,\delta}{\rho_{\gamma\delta} \sum_{\alpha,\beta}^n{ u_\alpha u_\beta^* A_\alpha^{\gamma \delta}(x_i) A_\beta^{\gamma \delta *}(x_i) } } } }

where we take n coherent amplitudes and allow incoherent sums indexed by γ, δ to allow for applications like spin-density matrices (ρ). When amplitude analysis fits contain amplitudes with not free parameters, it is convenient to rearrange the summations above, to pre-compute the sum over the intensities of the events:

Failed to parse (MathML with SVG or PNG fallback (recommended for modern browsers and accessibility tools): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle I=\sum_{\gamma,\delta}{\rho_{\gamma\delta} \sum_{\alpha,\beta}^n{ u_\alpha u_\beta^* \left[ \frac{1}{N_{gen}}\sum_i^N{ A_\alpha^{\gamma \delta}(x_i) A_\beta^{\gamma \delta *}(x_i) } \right] } } = \sum_{\gamma,\delta}{\rho_{\gamma\delta} \sum_{\alpha,\beta}^n{ u_\alpha u_\beta^* J^{\gamma\delta}_{\alpha \beta} } } }

storing the term in square brackets, a matrix indexed by α,β, for contractions with varying free production parameters u in the course of a fit.

When considering the uncertainty on the overall integral, both the errors on u parameters and those from the finite MC set of events will contribute. A single detected event (i) can be viewed as one sample in a process of independent events, having therefore a count uncertainty of σi=1. An integral over such events is then a weighted sum of such samples, having resulting in a contribution to the variance:

Failed to parse (MathML with SVG or PNG fallback (recommended for modern browsers and accessibility tools): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle \sigma_{MC}^2= \sum_i^N{ \sigma_i^2 \left| \frac{1}{N_{gen}} \sum_{\gamma,\delta}{\rho_{\gamma\delta} \sum_{\alpha,\beta}^n{ u_\alpha u_\beta^* A_\alpha^{\gamma \delta}(x_i) A_\beta^{\gamma \delta *}(x_i) } } \right|^2 } }

Failed to parse (MathML with SVG or PNG fallback (recommended for modern browsers and accessibility tools): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle = \sum_{\gamma,\delta,\gamma',\delta'}{ \rho_{\gamma\delta}\rho_{\gamma'\delta'} \sum_{\alpha,\beta,\alpha',\beta'}^n{ u_\alpha u_\beta^* u_{\alpha'} u_{\beta'}^* \left[ \frac{1}{N_{gen}^2} \sum_i^N{ A_\alpha^{\gamma \delta}(x_i) A_\beta^{\gamma \delta *}(x_i) A_{\alpha'}^{\gamma' \delta'}(x_i) A_{\beta'}^{\gamma' \delta' *}(x_i) } \right] } } }

The relevant piece to pre-compute over the event set for error calculation is shown in brackets. Turning our attention now to the contribution to error on the production parameters u:

Failed to parse (MathML with SVG or PNG fallback (recommended for modern browsers and accessibility tools): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle \sigma_{fit}^2= \left| \sum_k^n{ \sigma_{u_k} \frac{\partial}{\partial u_k}\left( \sum_{\gamma,\delta}{ \rho_{\gamma\delta} \sum_{\alpha,\beta}^n{u_\alpha u_\beta^* J_{\alpha\beta}} } \right) } \right|^2 }

Failed to parse (MathML with SVG or PNG fallback (recommended for modern browsers and accessibility tools): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle =\left( \sum_k^n{ \sigma_{u_k} \sum_{\gamma,\delta}{ \rho_{\gamma\delta} \sum_{\alpha,\beta}^n{\delta_{k\alpha} u_\beta^* J^{\gamma\delta}_{\alpha\beta}} } }\right) \left( \sum_{k'}^n{ \sigma^*_{u_{k'}} \sum_{\gamma',\delta'}{ \rho_{\gamma'\delta'} \sum_{\alpha',\beta'}^n{\delta_{k'\alpha'} u_{\beta'} J^{\gamma'\delta'*}_{\alpha'\beta'}} } }\right) }
Failed to parse (MathML with SVG or PNG fallback (recommended for modern browsers and accessibility tools): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle = \sum_{\gamma,\delta,\gamma',\delta'}{ \rho_{\gamma\delta} \rho_{\gamma'\delta'} \sum_{\alpha,\beta,\alpha',\beta'}^n{ \left(\sigma_{u_\alpha}\sigma^*_{u_{\alpha'}}\right) \left(u_\beta^* J^{\gamma\delta}_{\alpha\beta}\right) \left(u_{\beta'} J^{\gamma'\delta'*}_{\alpha'\beta'} \right) } } }
Failed to parse (MathML with SVG or PNG fallback (recommended for modern browsers and accessibility tools): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle = \sum_{\gamma,\delta,\gamma',\delta'}{ \rho_{\gamma\delta} \rho_{\gamma'\delta'} \sum_{\alpha,\alpha'}^n{ \left(\sigma_{u_\alpha}\sigma^*_{u_{\alpha'}}\right) G^{\gamma\delta*}_\alpha G_{\alpha'}^{\gamma\delta} } } }

The product of σ terms in the summation is represented by the error matrix derived from the fit. G was defined as

Failed to parse (MathML with SVG or PNG fallback (recommended for modern browsers and accessibility tools): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle G_\alpha^{\gamma\delta}=\sum_\beta{ u_\beta J_{\alpha\beta}^{\gamma\delta}} }

The overall uncertainty in the integral I defined in the beginning comes out to:

Failed to parse (MathML with SVG or PNG fallback (recommended for modern browsers and accessibility tools): Invalid response ("Math extension cannot connect to Restbase.") from server "https://wikimedia.org/api/rest_v1/":): {\displaystyle \sigma_I=\sqrt{\sigma^2_{MC} + \sigma^2_{fit}} }