From 0cf412e0b72db694ea67f096932062e63d17ee2c Mon Sep 17 00:00:00 2001 From: Jean Yang Date: Wed, 3 Dec 2008 23:36:51 -0500 Subject: [PATCH] added --- writeup/results_desc.tex | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/writeup/results_desc.tex b/writeup/results_desc.tex index 46952a8..95c5ded 100644 --- a/writeup/results_desc.tex +++ b/writeup/results_desc.tex @@ -1,7 +1,7 @@ \subsection{Call sequence analysis} We profiled how the code interacted with \texttt{ArrayList} library methods \texttt{add}, \texttt{get}, \texttt{remove}, \texttt{size}, \texttt{clear}, \texttt{toArray}, \texttt{iterator}, and \texttt{clone}. There were the functions that appeared in the libraries we profiled; there are other functions such as \texttt{addAll} that did not appear. We clustered a total of 3522 sequences. To provide an idea of what the sequences look like, the most frequently occurring sequences ares shown in Table~\ref{fig:freq_seq}. -Figure~ +Figure~\ref{fig:cluster3} shows the top transitions for different sequences. An interesting thing we found was that the length of the sequence does seem to correspond to its classification. Figure 3 shows the top sequences for each cluster for $k=3$. For $k = 2, 4, 5$, the largest cluster is associated with the shortest average sequence length (Figure~\ref{fig:transition_stats}. -- 2.11.4.GIT