Create RWD Documentation (#1699)

doc/user_manual/tsa.tex

@@ -115,10 +115,31 @@
     \end{itemize}
     \xmlNode{ARMA} further has the following subnodes:
     \begin{itemize}
-      \item \xmlNode{SignalLag}, \xmlDesc{integer, required field}, number of signal lag terms to
-        include in the autoregression term.
-      \item \xmlNode{NoiseLag}, \xmlDesc{integer, required field}, number of noise lag terms to
-        include in the moving average term.
+    \item \xmlNode{SignalLag}, \xmlDesc{integer, required field}, number of signal
+      lag terms to include in the autoregression term.
+    \item \xmlNode{NoiseLag}, \xmlDesc{integer, required field}, number of noise
+      lag terms to include in the moving average term.
     \end{itemize}
+    \item \xmlNode{RWD} characterizes the signal using Randomized Window
+      Decomposition. This algorithm leverages the \xmlNode{PostProcessor} subtype
+      \xmlAttr{TSACharacterizer}. This algorithm can only characterize, it cannot
+      generate data.
+      \xmlNode{RWD} has the following parameters:
+      \begin{itemize}
+      \item \xmlNode{signatureWindowLength} \xmlDesc{integer, required field} the
+        size of signature window, which represents as a snapshot for a certain
+        time step; typically represented as $w$ in literature.
+      \item \xmlNode{featureIndex} \xmlDesc{integer, required field} Index used
+        for feature selection, which requires pre-analysis for now, will be
+        addresses via other non human work required method.
+      \item \xmlNode{sampleType} \xmlDesc{integer, required field} Determines the
+        type of sampling to perform indicated by an integer value given ranging from $0$ to $2$.
+        \begin{itemize}
+        \item $0 = \text{Sequential Sampling}$
+        \item $1 = \text{Random Sampling}$
+        \item $2 = \text{Piecewise Sampling}$
+        \end{itemize}
+        \item \xmlNode{seed} \xmlDesc{integer} Indicating the random seed.
+      \end{itemize}
 \end{itemize}
 }