aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authornunzip <np.scarh@gmail.com>2018-11-17 13:11:46 +0000
committernunzip <np.scarh@gmail.com>2018-11-17 13:11:46 +0000
commitda362d8fe6f415aca3811f67beade3929e53dafe (patch)
tree3a5bf12050f47b560d9415068f4a9b7740007041
parent5e65497d7109db22137b9c27eb6262e612462fbc (diff)
downloadvz215_np1915-da362d8fe6f415aca3811f67beade3929e53dafe.tar.gz
vz215_np1915-da362d8fe6f415aca3811f67beade3929e53dafe.tar.bz2
vz215_np1915-da362d8fe6f415aca3811f67beade3929e53dafe.zip
Insert Scatter Matrix small description
-rwxr-xr-xreport/paper.md5
1 files changed, 4 insertions, 1 deletions
diff --git a/report/paper.md b/report/paper.md
index 253b1ba..e756f69 100755
--- a/report/paper.md
+++ b/report/paper.md
@@ -332,7 +332,10 @@ vaying between 0.11s(low M_pca) and 0.19s(high M_pca).
\end{center}
\end{figure}
-ADD RANK OF SCATTER MATRICES
+The scatter matrices obtained, S\textsubscript{B}(scatter matrix between classes) and
+S\textsubscript{W}(within-class scatter matrix), respectively show ranks of at most c-1(51) and
+N-c(312 maximum for our standard 70-30 split). The rank of S\textsubscript{W}
+will have the same value of M_pca for M_pca$\leq$N-c.
Testing with M_lda=50 and M_pca=115 gives 92.9% accuracy. The results of such test can be
observed in the confusion matrix shown in figure \ref{ldapca_cm}.