this is for holding javascript data
adam greenberg edited method.tex
about 10 years ago
Commit id: b859fc650fd2e8d1fee98f73226e93fd997dd6af
deletions | additions
diff --git a/method.tex b/method.tex
index e9cfe19..d32a5dd 100644
--- a/method.tex
+++ b/method.tex
...
\setcounter{section}{4}
\section{Method}
\subsection{The Square Root Information Filter}
The Square Root Information Filter (SRIF) was originally developed by Bierman in 1977 [ref]. The algorithm minimizes $\chi^2$ for time series data with Gaussian errors, and was based on the Kalman filter algorithm. SRIF is more stable, more accurate, and faster than the current algorithm used in \textbf{shape}. SRIF is also more numerically stable (and, in some cases faster) than a standard Jacobi-based steepest descent $\chi^2$-minimzation
routine (SDR). routine. My implmentation of SRIF includes some changes to the original algorithm, which will be discussed in section blah.blah. \par
The fundamental difference between SRIF algorithm and a classic SDR is the use of matrix square roots and Householder operations for increased numerical stability.
\subsection{Steepest Descent Routine}
A classical steepest descent routine (SDR) minimizes the weighted residuals between a model and data with Gaussian noise by determining the direction in parameter space in which the $\chi^2$ is decreasing fastest. Specifically, if one has a set of observables, $\vec{z}$ , and a model $M(x_1,x_2,...,x_n)$
The \emph{characteristic polynomial} $\chi(\lambda)$ of the
$3 \times 3$~matrix